From 6fafdacaf3e7ee9a1a8577a6cf056b9b477925b7 Mon Sep 17 00:00:00 2001
From: weidongfeng <2629144335@qq.com>
Date: Tue, 13 May 2025 01:57:55 +0000
Subject: [PATCH 1/2] =?UTF-8?q?!55=20!54=20=E6=96=B0=E5=A2=9E=E4=BA=86?=
=?UTF-8?q?=E5=9B=BE=E5=83=8F=E5=88=86=E5=89=B2=E5=92=8C=E8=BD=A6=E7=89=8C?=
=?UTF-8?q?=E8=AF=86=E5=88=AB=E6=A1=88=E4=BE=8B=20*=20Merge=20branch=20'ma?=
=?UTF-8?q?ster'=20of=20gitee.com:LockzhinerAI/LockzhinerVisionModule?=
=?UTF-8?q?=E2=80=A6=20*=20!54=20=E6=96=B0=E5=A2=9E=E4=BA=86=E5=9B=BE?=
=?UTF-8?q?=E5=83=8F=E5=88=86=E5=89=B2=E5=92=8C=E8=BD=A6=E7=89=8C=E8=AF=86?=
=?UTF-8?q?=E5=88=AB=E6=A1=88=E4=BE=8B=20*=20!53=20=E6=9B=B4=E6=94=B9?=
=?UTF-8?q?=E5=9B=BE=E6=A0=87=20*=20!50=20=E4=BF=AE=E6=94=B9=E4=B8=BB?=
=?UTF-8?q?=E6=96=87=E6=A1=A3=20*=20!49=20=E4=BF=AE=E6=94=B9=E4=BA=86?=
=?UTF-8?q?=E4=B8=BB=E7=95=8C=E9=9D=A2=E7=9A=84=E6=96=87=E6=A1=A3=20*=20!4?=
=?UTF-8?q?7=20=E4=BF=AE=E6=94=B9docker=E7=9A=84=E7=8E=AF=E5=A2=83?=
=?UTF-8?q?=E6=90=AD=E5=BB=BA=E6=95=99=E7=A8=8B=EF=BC=8C=E4=B8=BA=E5=8F=AF?=
=?UTF-8?q?=E8=83=BD=E5=87=BA=E7=8E=B0=E7=9A=84=E9=97=AE=E9=A2=98=E7=BB=99?=
=?UTF-8?q?=E5=87=BA=E4=BA=86=E8=A7=A3=E5=86=B3=E6=96=B9=E6=A1=88=E3=80=82?=
=?UTF-8?q?=20*=20!45=20=E6=9B=B4=E6=96=B0c++=E6=95=99=E7=A8=8B=E6=96=87?=
=?UTF-8?q?=E6=A1=A3=20*=20!43=20!41=20=E5=AE=8C=E5=96=84=E4=BA=86?=
=?UTF-8?q?=E9=83=A8=E5=88=86=E6=96=87=E6=A1=A3=20*=20!42=20!41=20?=
=?UTF-8?q?=E5=AE=8C=E5=96=84=E4=BA=86=E9=83=A8=E5=88=86=E6=96=87=E6=A1=A3?=
=?UTF-8?q?=20*=20=E4=BF=AE=E5=A4=8D=E4=BA=BA=E8=84=B8=E8=AF=86=E5=88=AB?=
=?UTF-8?q?=E7=B3=BB=E7=BB=9F=E6=96=87=E6=A1=A3=20*=20!39=20=E6=B7=BB?=
=?UTF-8?q?=E5=8A=A0wifi=E9=85=8D=E7=BD=AE?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
---
.../README.md | 2 +-
Cpp_example/D08_pp_humanseg/CMakeLists.txt | 39 +
Cpp_example/D08_pp_humanseg/README.md | 450 ++++++
Cpp_example/D08_pp_humanseg/images/result.jpg | Bin 0 -> 212661 bytes
Cpp_example/D08_pp_humanseg/images/test.png | Bin 0 -> 1474571 bytes
Cpp_example/D08_pp_humanseg/images/view.png | Bin 0 -> 161901 bytes
Cpp_example/D08_pp_humanseg/pp_humanseg.cc | 232 +++
.../D09_plate_recognize/CMakeLists.txt | 44 +
Cpp_example/D09_plate_recognize/README.md | 582 ++++++++
.../D09_plate_recognize/images/D09_1.png | Bin 0 -> 230032 bytes
.../D09_plate_recognize/images/D09_2.jpg | Bin 0 -> 148982 bytes
.../D09_plate_recognize/images/D09_3.png | Bin 0 -> 26187 bytes
.../D09_plate_recognize/images/D09_4.png | Bin 0 -> 1347373 bytes
Cpp_example/D09_plate_recognize/myfontface.h | 1323 +++++++++++++++++
.../D09_plate_recognize/plate_recognize.cc | 251 ++++
README.md | 18 +-
16 files changed, 2938 insertions(+), 3 deletions(-)
create mode 100644 Cpp_example/D08_pp_humanseg/CMakeLists.txt
create mode 100644 Cpp_example/D08_pp_humanseg/README.md
create mode 100644 Cpp_example/D08_pp_humanseg/images/result.jpg
create mode 100644 Cpp_example/D08_pp_humanseg/images/test.png
create mode 100755 Cpp_example/D08_pp_humanseg/images/view.png
create mode 100644 Cpp_example/D08_pp_humanseg/pp_humanseg.cc
create mode 100644 Cpp_example/D09_plate_recognize/CMakeLists.txt
create mode 100644 Cpp_example/D09_plate_recognize/README.md
create mode 100755 Cpp_example/D09_plate_recognize/images/D09_1.png
create mode 100755 Cpp_example/D09_plate_recognize/images/D09_2.jpg
create mode 100755 Cpp_example/D09_plate_recognize/images/D09_3.png
create mode 100755 Cpp_example/D09_plate_recognize/images/D09_4.png
create mode 100755 Cpp_example/D09_plate_recognize/myfontface.h
create mode 100644 Cpp_example/D09_plate_recognize/plate_recognize.cc
diff --git a/Cpp_example/B02_Image_information_statistics/README.md b/Cpp_example/B02_Image_information_statistics/README.md
index 124505f..aba6ccc 100755
--- a/Cpp_example/B02_Image_information_statistics/README.md
+++ b/Cpp_example/B02_Image_information_statistics/README.md
@@ -105,7 +105,7 @@ cv::minMaxLoc(src, &minVal, &maxVal, &minLoc, &maxLoc);
### 3.1 流程图
-
+
### 3.2 代码解释
- 读取图像文件
diff --git a/Cpp_example/D08_pp_humanseg/CMakeLists.txt b/Cpp_example/D08_pp_humanseg/CMakeLists.txt
new file mode 100644
index 0000000..d9e44bb
--- /dev/null
+++ b/Cpp_example/D08_pp_humanseg/CMakeLists.txt
@@ -0,0 +1,39 @@
+cmake_minimum_required(VERSION 3.10)
+
+project(pp_humanseg)
+
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+
+# 定义项目根目录路径
+set(PROJECT_ROOT_PATH "${CMAKE_CURRENT_SOURCE_DIR}/../..")
+message("PROJECT_ROOT_PATH = " ${PROJECT_ROOT_PATH})
+
+include("${PROJECT_ROOT_PATH}/toolchains/arm-rockchip830-linux-uclibcgnueabihf.toolchain.cmake")
+
+# 定义 OpenCV SDK 路径
+set(OpenCV_ROOT_PATH "${PROJECT_ROOT_PATH}/third_party/opencv-mobile-4.10.0-lockzhiner-vision-module")
+set(OpenCV_DIR "${OpenCV_ROOT_PATH}/lib/cmake/opencv4")
+find_package(OpenCV REQUIRED)
+set(OPENCV_LIBRARIES "${OpenCV_LIBS}")
+
+# 定义 LockzhinerVisionModule SDK 路径
+set(LockzhinerVisionModule_ROOT_PATH "${PROJECT_ROOT_PATH}/third_party/lockzhiner_vision_module_sdk")
+set(LockzhinerVisionModule_DIR "${LockzhinerVisionModule_ROOT_PATH}/lib/cmake/lockzhiner_vision_module")
+find_package(LockzhinerVisionModule REQUIRED)
+
+# 配置rknpu2
+set(RKNPU2_BACKEND_BASE_DIR "${LockzhinerVisionModule_ROOT_PATH}/include/lockzhiner_vision_module/vision/deep_learning/runtime")
+if(NOT EXISTS ${RKNPU2_BACKEND_BASE_DIR})
+ message(FATAL_ERROR "RKNPU2 backend base dir missing: ${RKNPU2_BACKEND_BASE_DIR}")
+endif()
+
+
+add_executable(Test-pp_humanseg pp_humanseg.cc)
+target_include_directories(Test-pp_humanseg PRIVATE ${LOCKZHINER_VISION_MODULE_INCLUDE_DIRS} ${rknpu2_INCLUDE_DIRS} ${RKNPU2_BACKEND_BASE_DIR})
+target_link_libraries(Test-pp_humanseg PRIVATE ${OPENCV_LIBRARIES} ${NCNN_LIBRARIES} ${LOCKZHINER_VISION_MODULE_LIBRARIES})
+
+install(
+ TARGETS Test-pp_humanseg
+ RUNTIME DESTINATION .
+)
\ No newline at end of file
diff --git a/Cpp_example/D08_pp_humanseg/README.md b/Cpp_example/D08_pp_humanseg/README.md
new file mode 100644
index 0000000..7deb130
--- /dev/null
+++ b/Cpp_example/D08_pp_humanseg/README.md
@@ -0,0 +1,450 @@
+# 图像分割
+本章节在 Lockzhiner Vision Module 上基于paddleseg模型, 实现了一个PP-HumanSeg人像分割案例。
+## 1. 基本知识简介
+### 1.1 人像分割简介
+人像分割是一种基于计算机视觉的技术，通过深度学习算法精准识别图像或视频中的人物主体，将其与背景进行像素级分离。该技术可实时运行于移动端及嵌入式设备，广泛应用于虚拟背景、智能抠图、视频会议美颜等场景，支持复杂光照、多样姿态和遮挡情况下的高精度分割，兼顾处理速度与效果。
+### 1.2 人像分割常用方法
+目前对于实现人像分割任务的方法有很多，下面介绍几种常用的人像分割实现方法。
+- 传统算法（如GrabCut）：基于颜色直方图与图割优化，适合简单背景，计算量小但精度有限。
+- U-Net系列：编码器-解码器结构，医学图像起家，适合精细边缘，需较高算力。
+- DeepLab系列：采用空洞卷积扩大感受野，擅长复杂场景，模型较大。
+- BiSeNet：双分支结构平衡速度与精度，实时分割首选，移动端友好。
+- PP-HumanSeg：百度自研轻量模型，专为人像优化，支持半监督训练，RKNN部署效率高。
+
+这些方法各有优势，其中在工业部署方面PP-HumanSeg（精度与速度平衡）和BiSeNet（高性价比）更适合，可配合OpenCV后处理优化边缘。
+
+---
+
+## 2. C++ API 文档
+### 2.1 RKNPU2Backend 类
+#### 2.1.1 头文件
+```cpp
+#include "rknpu2_backend/rknpu2_backend.h"
+```
+- 作用：创建一个RKNPU2Backend类，用于实现对rknn模型的处理。
+
+#### 2.1.2 构造类函数
+```cpp
+ockzhiner_vision_module::vision::RKNPU2Backend backend;
+```
+- 作用：创建一个RKNPU2Backen类型的对象实例，用于实现人像分割。
+- 参数说明：
+ - 无
+- 返回值：
+ - 无
+
+#### 2.1.3 Initialize 函数
+```cpp
+bool Initialize(const std::string &model_path, const std::string ¶m_path = "") override;
+```
+- 作用：初始化 RKNN 模型，加载模型文件和可选参数文件，完成推理引擎的准备工作。
+- 参数说明：
+ - model_path：必需参数，RKNN 模型文件路径（.rknn 格式）。
+ - param_path：可选参数，额外参数文件路径（某些场景下用于补充模型配置，默认空字符串）。
+- 返回值：返回ture/false，表示模型初始化是否成功。
+
+#### 2.1.4 Run 函数
+```cpp
+bool Run();
+```
+- 作用：执行模型推理计算，驱动输入数据通过模型计算得到输出结果。
+- 参数说明：
+ - 无
+- 返回值：
+ - true：推理执行成功。
+ - false：推理失败（可能原因：输入数据未准备、内存不足等）。
+
+#### 2.1.5 GetInputAttrs 函数
+```cpp
+const std::vector& GetInputAttrs() const;
+```
+- 作用：获取模型所有输入张量的属性信息（维度/形状、数据类型、量化参数等）。
+- 参数说明：
+ - 无
+- 返回值：常量引用形式的 rknn_tensor_attr 向量，包含输入张量属性。
+
+#### 2.1.6 GetOutputAttrs 函数
+```cpp
+const std::vector& GetInputMemories() const;
+```
+- 作用：获取模型所有输出张量的属性信息。
+- 参数说明：无
+- 返回值：常量引用形式的 rknn_tensor_attr 向量，包含输出张量属性。
+
+## 3. PP-Humanseg人像分割代码解析
+### 3.1 流程图
+
+
+
+### 3.2 核心代码解析
+- 初始化模型
+```cpp
+backend.Initialize(model_path)
+```
+- 获取输入输出属性
+```cpp
+const auto& input_attrs = backend.GetInputAttrs();
+const auto& output_attrs = backend.GetOutputAttrs();
+```
+- 对输入图像进行推理
+```cpp
+backend.Run()
+```
+自定义函数说明
+- pp-humanseg输入预处理
+```cpp
+cv::Mat preprocess(const cv::Mat& image, const std::vector& input_dims)
+```
+- 作用：对输入图像进行预处理操作，包括 尺寸调整、颜色空间转换 和 量化处理，使其符合 RKNN 模型的输入要求。
+- 参数说明：
+ - image：输入图像（BGR 格式的 cv::Mat 对象）。
+ - input_dims：模型输入张量的维度定义（需满足 [1, H, W, 3] 的 NHWC 格式）。
+- 返回值：
+ - 返回预处理后的量化张量（cv::Mat，数据类型为 CV_8S）。
+ - 若输入维度不合法，返回空矩阵（cv::Mat()）并报错。
+- pp-humanseg输入后处理
+```cpp
+cv::Mat postprocess(const rknn_tensor_mem* output_mem,
+ const std::vector& output_dims,
+ const cv::Size& target_size)
+```
+- 作用：将模型输出的原始张量转换为高精度分割掩膜，包含 概率解码、动态阈值分割、形态学优化和边缘增强等步骤，最终生成与原始图像尺寸匹配的二值化掩膜。
+- 参数说明：
+ - output_mem：模型输出的内存指针，包含量化后的原始数据。
+ - output_dims：模型输出的维度信息，需满足 [1, 2, H, W] 的 NCHW 格式。
+ - target_size：目标输出尺寸。
+- 返回值：返回优化后的二值化掩膜。
+
+### 3.3 完整代码实现
+```cpp
+#include
+#include
+#include
+#include "rknpu2_backend/rknpu2_backend.h"
+#include
+#include
+#include
+
+using namespace std::chrono;
+// 输入和输出量化的scale和zeropoint
+const float INPUT_SCALE = 1.0f / 255.0f;
+const int INPUT_ZP = -128;
+const float OUTPUT_SCALE = 0.034558f;
+const int OUTPUT_ZP = -128;
+
+// 预处理函数
+cv::Mat preprocess(const cv::Mat& image, const std::vector& input_dims) {
+ // 确保输入维度为NHWC [1, H, W, 3]
+ if (input_dims.size() != 4 || input_dims[0] != 1 || input_dims[3] != 3) {
+ std::cerr << "Invalid input dimensions" << std::endl;
+ return cv::Mat();
+ }
+
+ const size_t input_h = input_dims[1];
+ const size_t input_w = input_dims[2];
+
+ // Resize并转换颜色空间
+ cv::Mat resized, rgb;
+ cv::resize(image, resized, cv::Size(input_w, input_h));
+ cv::cvtColor(resized, rgb, cv::COLOR_BGR2RGB);
+
+ // 量化到INT8
+ cv::Mat quantized;
+ rgb.convertTo(quantized, CV_8S, 255.0 * INPUT_SCALE, INPUT_ZP);
+
+ return quantized;
+}
+
+// 后处理函数
+cv::Mat postprocess(const rknn_tensor_mem* output_mem,
+ const std::vector& output_dims,
+ const cv::Size& target_size) {
+ // 验证输出维度为NCHW [1, 2, H, W]
+ if (output_dims.size() != 4 || output_dims[0] != 1 || output_dims[1] != 2) {
+ std::cerr << "Invalid output dimensions" << std::endl;
+ return cv::Mat();
+ }
+
+ const int8_t* data = static_cast(output_mem->virt_addr);
+ const int h = output_dims[2];
+ const int w = output_dims[3];
+
+ // ================= 1. 概率图生成优化 =================
+ cv::Mat prob_map(h, w, CV_32FC1);
+ // 基于192x192模型的缩放补偿
+ float spatial_weight = 1.0f - (h * w) / (192.0f * 192.0f);
+
+ for (int y = 0; y < h; ++y) {
+ for (int x = 0; x < w; ++x) {
+ const int bg_idx = 0 * h * w + y * w + x;
+ const int fg_idx = 1 * h * w + y * w + x;
+ // 带饱和保护的反量化
+ float bg_logit = std::clamp((data[bg_idx] - OUTPUT_ZP) * OUTPUT_SCALE, -10.0f, 10.0f);
+ float fg_logit = std::clamp((data[fg_idx] - OUTPUT_ZP) * OUTPUT_SCALE, -10.0f, 10.0f);
+
+ // 空间注意力加权（中心区域增强）
+ float center_weight = 1.0f - (std::abs(x - w/2.0f)/(w/2.0f) + std::abs(y - h/2.0f)/(h/2.0f))/2.0f;
+ fg_logit *= (1.2f + 0.3f * center_weight * spatial_weight);
+
+ // 稳健的Softmax计算
+ float max_logit = std::max(bg_logit, fg_logit);
+ float exp_sum = expf(bg_logit - max_logit) + expf(fg_logit - max_logit);
+ prob_map.at(y, x) = expf(fg_logit - max_logit) / (exp_sum + 1e-8f);
+ }
+ }
+
+ // ================= 2. 自适应阈值优化 =================
+ cv::Mat binary_mask;
+ // 重点区域检测
+ cv::Mat prob_roi = prob_map(cv::Rect(w/4, h/4, w/2, h/2));
+ float center_mean = cv::mean(prob_roi)[0];
+ float dynamic_thresh = std::clamp(0.45f - (center_mean - 0.5f) * 0.3f, 0.25f, 0.65f);
+
+ cv::threshold(prob_map, binary_mask, dynamic_thresh, 255, cv::THRESH_BINARY);
+ binary_mask.convertTo(binary_mask, CV_8U);
+
+ // ================= 3. 多尺度形态学处理 =================
+ std::vector mask_pyramid;
+ // 构建金字塔
+ cv::buildPyramid(binary_mask, mask_pyramid, 2);
+
+ // 小尺度去噪
+ cv::Mat kernel1 = cv::getStructuringElement(cv::MORPH_ELLIPSE, cv::Size(3,3));
+ cv::morphologyEx(mask_pyramid[1], mask_pyramid[1], cv::MORPH_OPEN, kernel1);
+
+ // 中尺度填充
+ cv::Mat kernel2 = cv::getStructuringElement(cv::MORPH_ELLIPSE, cv::Size(5,5));
+ cv::morphologyEx(mask_pyramid[0], mask_pyramid[0], cv::MORPH_CLOSE, kernel2);
+
+ // 金字塔重建
+ cv::Mat refined_mask;
+ cv::pyrUp(mask_pyramid[1], refined_mask, mask_pyramid[0].size());
+ cv::bitwise_and(refined_mask, mask_pyramid[0], refined_mask);
+
+ // ================= 4. 智能边缘优化 =================
+ cv::Mat edge_weights;
+ cv::distanceTransform(refined_mask, edge_weights, cv::DIST_L2, 3);
+ cv::normalize(edge_weights, edge_weights, 0, 1.0, cv::NORM_MINMAX);
+
+ cv::Mat probabilistic_edges;
+ cv::Canny(refined_mask, probabilistic_edges, 50, 150);
+ probabilistic_edges.convertTo(probabilistic_edges, CV_32F, 1.0/255.0);
+
+ cv::Mat final_edges;
+ cv::multiply(probabilistic_edges, edge_weights, final_edges);
+ final_edges.convertTo(final_edges, CV_8U, 255.0);
+
+ // ================= 5. 多模态融合输出 =================
+ cv::Mat resized_mask;
+ cv::resize(refined_mask, resized_mask, target_size, 0, 0, cv::INTER_LANCZOS4);
+
+ cv::Mat final_mask;
+ cv::bilateralFilter(resized_mask, final_mask, 5, 15, 15);
+
+ // 最终保边处理
+ cv::Mat edge_mask_hr;
+ cv::resize(final_edges, edge_mask_hr, target_size, 0, 0, cv::INTER_NEAREST);
+ final_mask.setTo(255, edge_mask_hr > 128);
+
+ return final_mask;
+}
+
+int main(int argc, char* argv[]) {
+ if (argc != 3) {
+ std::cerr << "Usage: " << argv[0] << " " << std::endl;
+ return 1;
+ }
+
+ const std::string model_path = argv[1];
+ const std::string image_path = argv[2];
+
+ // 初始化RKNN后端
+ lockzhiner_vision_module::vision::RKNPU2Backend backend;
+ if (!backend.Initialize(model_path)) {
+ std::cerr << "Failed to initialize RKNN backend" << std::endl;
+ return -1;
+ }
+
+ // 加载图像
+ cv::Mat image = cv::imread(image_path);
+ if (image.empty()) {
+ std::cerr << "Failed to read image: " << image_path << std::endl;
+ return -1;
+ }
+
+ // 获取输入属性
+ const auto& input_attrs = backend.GetInputAttrs();
+ if (input_attrs.empty()) {
+ std::cerr << "No input attributes found" << std::endl;
+ return -1;
+ }
+ const auto& input_attr = input_attrs[0];
+ std::vector input_dims(input_attr.dims, input_attr.dims + input_attr.n_dims);
+
+ // 预处理
+ cv::Mat preprocessed = preprocess(image, input_dims);
+ if (preprocessed.empty()) {
+ std::cerr << "Preprocessing failed" << std::endl;
+ return -1;
+ }
+
+ // 验证输入数据尺寸
+ const size_t expected_input_size = input_attr.size_with_stride;
+ const size_t actual_input_size = preprocessed.total() * preprocessed.elemSize();
+ if (expected_input_size != actual_input_size) {
+ std::cerr << "Input size mismatch! Expected: " << expected_input_size
+ << ", Actual: " << actual_input_size << std::endl;
+ return -1;
+ }
+
+ // 拷贝输入数据
+ const auto& input_memories = backend.GetInputMemories();
+ if (input_memories.empty() || !input_memories[0]) {
+ std::cerr << "Invalid input memory" << std::endl;
+ return -1;
+ }
+ memcpy(input_memories[0]->virt_addr, preprocessed.data, actual_input_size);
+
+ // 执行推理
+ high_resolution_clock::time_point start_time =
+ high_resolution_clock::now();
+ if (!backend.Run()) {
+ std::cerr << "Inference failed" << std::endl;
+ return -1;
+ }
+
+ // 获取输出
+ const auto& output_attrs = backend.GetOutputAttrs();
+ if (output_attrs.empty()) {
+ std::cerr << "No output attributes found" << std::endl;
+ return -1;
+ }
+ const auto& output_memories = backend.GetOutputMemories();
+ if (output_memories.empty() || !output_memories[0]) {
+ std::cerr << "Invalid output memory" << std::endl;
+ return -1;
+ }
+
+ // 后处理
+ const auto& output_attr = output_attrs[0];
+ std::vector output_dims(output_attr.dims, output_attr.dims + output_attr.n_dims);
+ cv::Mat mask = postprocess(output_memories[0], output_dims, image.size());
+ high_resolution_clock::time_point end_time = high_resolution_clock::now();
+ auto time_span = duration_cast(end_time - start_time);
+ std::cout << "单张图片推理时间(ms): " << time_span.count() << std::endl;
+ // 生成结果
+ cv::Mat result;
+ cv::bitwise_and(image, image, result, mask);
+
+ // 保存结果
+ const std::string output_path = "result.jpg";
+ cv::imwrite(output_path, result);
+ std::cout << "Result saved to: " << output_path << std::endl;
+
+ // 显示调试视图
+ cv::imshow("Original", image);
+ cv::imshow("Mask", mask);
+ cv::imshow("Result", result);
+ cv::waitKey(0);
+
+ return 0;
+}
+```
+
+---
+
+## 4. 编译调试
+### 4.1 编译环境搭建
+- 请确保你已经按照 [开发环境搭建指南](../../../../docs/introductory_tutorial/cpp_development_environment.md) 正确配置了开发环境。
+- 同时已经正确连接开发板。
+### 4.2 Cmake介绍
+```cmake
+cmake_minimum_required(VERSION 3.10)
+
+project(pp_humanseg)
+
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+
+# 定义项目根目录路径
+set(PROJECT_ROOT_PATH "${CMAKE_CURRENT_SOURCE_DIR}/../..")
+message("PROJECT_ROOT_PATH = " ${PROJECT_ROOT_PATH})
+
+include("${PROJECT_ROOT_PATH}/toolchains/arm-rockchip830-linux-uclibcgnueabihf.toolchain.cmake")
+
+# 定义 OpenCV SDK 路径
+set(OpenCV_ROOT_PATH "${PROJECT_ROOT_PATH}/third_party/opencv-mobile-4.10.0-lockzhiner-vision-module")
+set(OpenCV_DIR "${OpenCV_ROOT_PATH}/lib/cmake/opencv4")
+find_package(OpenCV REQUIRED)
+set(OPENCV_LIBRARIES "${OpenCV_LIBS}")
+
+# 定义 LockzhinerVisionModule SDK 路径
+set(LockzhinerVisionModule_ROOT_PATH "${PROJECT_ROOT_PATH}/third_party/lockzhiner_vision_module_sdk")
+set(LockzhinerVisionModule_DIR "${LockzhinerVisionModule_ROOT_PATH}/lib/cmake/lockzhiner_vision_module")
+find_package(LockzhinerVisionModule REQUIRED)
+
+# 配置rknpu2
+set(RKNPU2_BACKEND_BASE_DIR "${LockzhinerVisionModule_ROOT_PATH}/include/lockzhiner_vision_module/vision/deep_learning/runtime")
+if(NOT EXISTS ${RKNPU2_BACKEND_BASE_DIR})
+ message(FATAL_ERROR "RKNPU2 backend base dir missing: ${RKNPU2_BACKEND_BASE_DIR}")
+endif()
+
+
+add_executable(Test-pp_humanseg pp_humanseg.cc)
+target_include_directories(Test-pp_humanseg PRIVATE ${LOCKZHINER_VISION_MODULE_INCLUDE_DIRS} ${rknpu2_INCLUDE_DIRS} ${RKNPU2_BACKEND_BASE_DIR})
+target_link_libraries(Test-pp_humanseg PRIVATE ${OPENCV_LIBRARIES} ${NCNN_LIBRARIES} ${LOCKZHINER_VISION_MODULE_LIBRARIES})
+
+install(
+ TARGETS Test-pp_humanseg
+ RUNTIME DESTINATION .
+)
+```
+### 4.3 编译项目
+使用 Docker Destop 打开 LockzhinerVisionModule 容器并执行以下命令来编译项目。
+```bash
+# 进入Demo所在目录
+cd /LockzhinerVisionModuleWorkSpace/LockzhinerVisionModule/Cpp_example/D08_pp_humanseg
+# 创建编译目录
+rm -rf build && mkdir build && cd build
+# 配置交叉编译工具链
+export TOOLCHAIN_ROOT_PATH="/LockzhinerVisionModuleWorkSpace/arm-rockchip830-linux-uclibcgnueabihf"
+# 使用cmake配置项目
+cmake ..
+# 执行编译项目
+make -j8 && make install
+```
+
+在执行完上述命令后，会在build目录下生成可执行文件。
+
+---
+
+## 5. 执行结果
+### 5.1 运行前准备
+- 请确保你已经下载了 [凌智视觉模块人像分割模型](https://gitee.com/LockzhinerAI/LockzhinerVisionModule/releases/download/v0.0.6/pp-humanseg.rknn)
+### 5.2 运行过程
+```shell
+chmod 777 Test-pp_humanseg
+# 对人像进行分割
+./Test-pp_humanseg pp-humanseg.rknn image_path
+```
+### 5.3 运行效果
+#### 5.3.1 人像分割结果
+- 原始图像
+
+
+
+- 分割结果
+
+
+
+#### 5.3.2 注意事项
+由于分割的模型很小，并且在模型转换过程中会有精度损失，所以在测试的时候尽量选择背景比较纯净的图像效果比较好。
+
+---
+
+## 6. 总结
+通过上述内容，我们成功实现了一个简单的人像分割的例子，包括：
+
+- 加载图像分割的rknn模型和待分割图像。
+- 图像预处理和模型推理。
+- 图像后处理并保存分割结果。
\ No newline at end of file
diff --git a/Cpp_example/D08_pp_humanseg/images/result.jpg b/Cpp_example/D08_pp_humanseg/images/result.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..ad3ed101c345e5a60c99fc2c96be208b7e081a3a
GIT binary patch
literal 212661
zcmeFZbwE_#yDvNlh_oP`($dn+n1rYZNSD&xU4tkg9TJiv9V3!M58a`3cZe`B12POS
z40rpx=iYPAee0a}egFLKnf1?}z4ltqv)4TFc|L1j&t0#9Za-5|R{`PS;ep-&AJFw8
z=rIWY#*M#U;2;3LgtrI@2?z*Dh=^|9A|oLqBPAgvC8waiO-?~YK}vd??l#pO8d_Rf
zGD>;|IvNIQ8d{paH^IXP`VbHj6A}{BkduC;#}#Qx#P;
z^=CS-b@lWO42>+UtZi)V>>WJ4ynTHA`~%*92#<)2`WT&@@+mbfJ>zrc*Mh>L;*!#`
z@|xPZ`UY5IQ*&2$Pj6rUz~IotkH3?&%Tj8ecO}H9l
zebD^W5Pm#!4LTeGVcWmLU;T$u;Qz|+K5s(IwaiP2aJSME&MQ~X9U*7dd{{XdT!kpk
zX!UCTY3^Fc8!hvaYY=kl=%Nz-uM51|iLXInVi=L06-s*S)7ko(MdjhE7~7j!+m=38
zIM)T4-K{=pOGjHD6?$X5bMH0wn=9C-sM$zGzG4m6n*~)4dJCp?u;;@h8j=-E5~+F|
zk7dVbvTZ@Ldw9T&`J4x@LE*#KpfuPu$VXpm(-p%FjSR*0Zd`-jx692!72mEeUwq56
zg)T0bUDaHW*DRh@Gmui_?K31LON|*8B0l5F&gcxjtg|ldDghz)=2FW
z&z6vDkfiZJ-yMe4Y}j(gFn&5a@+n!~xey@RHR#M*0mqVa4bsiJ1_`_34yrWFF+#ST
zsr&1Bta+DiHP@i6W}RygVLV3a8uVy99(4^eqcg{*YrYKiPl0sa>AD7uT3>^Pjjus1
z_slVzA^v|wd3O!6fd7@HxC%=rvDObE!<+y&K?5DNDKSC`oo}x};{6F=q&d)c@Kfjs
zCC0<{O3(5d^x>}xW_PYZ39Zb5$o+mZE@W9G?5p)!rJduy>~6%iPWv@AM2q&eR7NynX9!eTx_V
zu|@QiDJmUH<=Rc(T4x9%-TAwx-e4MK7Dw}2Asm?Zow*$E4=(yX`txL8YE5%5gcw?`l=sQe
zR20xMCrzDwGfE_+(R%Pt`~O1+de@*(EKwEqB`P9y-Idf{pJ@eN{`_uTbpyUvT$nm{
zBCYOY35DC`4Hed;n8=UxOCT(DC1qS2HkI`zlPgc5^#?`tS|&W7@ReJ6bevZuDwD
zDwQs?sQAPZO376p4N9p5DC0BXj;&0gpBPX27eWHRbj9*qgD9jBq%PxtS+FqcEc0(#
zhC?!i(U!!qu46lwjf`ma4-lA47{m$cXn^rvtDvOw^V!){ZO*+08MY)dlg2P>hf&J^
z-0=U{WN7=|Wi9l^QOi9lEO0vCAVQ*iA3NOM?(B4A}g|C%8)$Es^krwrT{dPCYDYrXk|bbPeT;2Lzy<8K}Xw>FQ&kcoJ|FXARoqyclP{c%fquM0%Cul0jU54Pt(J4RUROVB?-%MDn9;vC0~2{a2)H
ziom3T)2~516Oit$5HiS*>dn?`&}=wxidTP3nei&|8e{{f5--Pe_a&v_(D^pt9B6<`
z(lv-+@+z;>b(oUxI$KGXbly;eF$xmyUXNxN(?|PjKSHoW%&=!^?_T&(DhPz+lZ7
zq7B^|d6&Fzeu@v_Be>q_T+KR{#kgybqag}r-!}^Pug$&hPpSjI%7LtX$tAcdQN9c5
zVqd2u!_)}=LU6+yo2L*5>oC|mk}_XGj}ea}CQwwrh%5=I=%*LJ009_LHpoUXY86KY
zD3rR53+;QA9>NsTe-YnOhVeLgJ>^Yq#sJ-Fo#lpQ+HtDbi1nbl!gc4$-VRvOK$dAI
zgC1NJNs3ac^150Qs)A*dBzl!I6FwpR;cH2G1kg^=jXYLL=wEd%J{xWZe5o$>R{hvD
zhz_%c_xFj5;+|BBMy)n%E}LjjWbw{Tt*N>1PkEX>EDxEwKo^HoBWm+njWdn-Y?O7U
zb5C4>DOX|cc(oLu%;TAH8`9?3kJ?pO>hv{xCw3al7G&uW;HHxVyA@?FHT@D)-FKE5
zEft-CZl+%8tM%J|0$I7YyczH6OrR}mlpMeqZg|HYc*g-Xa1Cnw`}8^WzheP^92V(E
zgmy+9VL5XJk*m?Wt;rJ#`+-mqQg!GdhsO}3hDimPyxIr(vY|UeiJC)3u86Xr5k1W>
zeU5?~L|DPRg@RWBhCLRtRtu**@M_HK$_1fB>PtU|4Z2p2x|VcRJ9B?emdS~6
zS&;sOPmt5iAPXh|{_EiRp=Fo2)3n-6$~vdxYf$~Eeys5(u%bRa{_iX5J;Sb&eiVEy
zlO5D*UE1qi+`nN?-0{8AW?XfecEe}*4u_1^$YuPR?3+fdJ`eRB5j`smAC4FZI>-sp
ztI=SVR^zINo0JSr(_23!pI(ZhHM5$+E9w_CIF2{(_i-(>Yo-mRgQRSNQ~>3P{(`Xu
z;qWkVkhe9Xtre71^F^~cUNW@H?v&}le4D`W536gKB-o#jAKb^K3zby;>F=A?Z8v*nh+&AVHmLCf0MmIf=Jqj!GBLto0MXw??FRSKu#=19MI}&c2
z|CT=s3%b(Op1>NQPx{?0IIOe#23a0haEWYzCAg~NCZ4>y2C0PfUxTV~A1^5}=6z>p
zI2qg`Zu;1zo?QEQK}mUHY-XO@n&)`}%r#3wF^=~>{kCZ#EDSfDS6>nVcgi8v#NZ)S
zc5&3EzI`%js}9rNr0Pxw&d@iuP3e^U>I6~F1)fpa48!5eyfn(`_>0#dOU=czZopo&
zmq35LSYJ%I98lff3H<;)9%JDq7w7+N-k9j~4K8#EVG%O_>;x
zR$Yj5IY!^OHB%!_WwRUR-jP1h_`oVz&N2z#T7xd`Rjt~=Wya3et3)`EI8cW4hp@Iz
z9iB?fw0c$=cHb-u(IRYXpYgSGCXQ9BP9AKn4MBD?J8L#KuGVQHq$neuOB!l(o;UV+
zcw%()mI}XlDC1)Sky&V_ydI;-@KbUP?P)0kOIRS(#n`_%Lej>PMU&;8u1Ts8@!gC9
z9KsLMJHBXXgd#ZZEK>XNINLv~X@8+rbc3&FRI@MOL#@e@1Q*LDcYOV>=Mw*zreuA6
z23mfVD(K}bR-Hd^RK*|P^F2rZcINg@ds0Sv3g0(PFw|zy%OJ5HU&0rhqa^A@|Kq3g
zbCCx-sZzuSr@b~RO-|^t;R6Pm)B4Z6hzPi6$#=VAYo&l$%Y3XGiY}(k*g;$sRwA{-
zXzP+Hzae&_Rdg6Rb+NiWj%eCYB2QL&WM0=Ld)_E@xS)xcEM8qKZA^M=!@5{PdtN%5
z^^hytDd0nRqtokovyqBC=ODMM3W{nc=em}!tc@!lVm%ypDDU_!zM9VWw`oq+b9rfS
zq0exUf;i7e(W*cDOnSrf8uYt-X79+!E
z85Sy)^k%-1l@xt}g5&)cj@{y0$q=UQWYh1Vno>$_{f7Z3Dh;)a#pzG27!-pY4l=W6
zP!aJe+O4OYP1m407q&m=-(ns;RQy?V7!g1pfs=4FBKx>)+D?O%VM4Fo3R%c(NZIaJ
zo_+ZtB?Y^`MJ2tIY)AhzfrtiB|5QRN)&x#^4RU(&bFojux5SYS+!Cob!5zz1$d3c)l&K#);4}();tqU<-W02c^F|Q`A;}83UjkPp+7k1>R4R&ta
z+7KEVJTr~H-7nzRzmpe^ee3^aMOBEJMqnNr7u<>DLRZyC43RoNGv}YeYrG1nGc|uu
zV><<}(?=kXyJ&M1UCI|s@wnNvRPdu8tE9nn6LN|f8Z}%B8bXdIHEJO39)IU-YtJ=h
zX~1FB+2a&1=|3n_`Z1d|qY|QEd-B1DF!z7ut>|8RPF|aMC$)1yj<^u5KP5H;$k*bd>wWYT`6syDV-gx6cfb~+wkz<)r
z*05+N;d;$G`+Ao%+R{Phjs@58mxz&^%rSsQ>f?
zJ#z*3I-1v3prqW(6vc9z$xR)s9Fhh`+ChcZl`T{iD{oqll&U6wpr$%TVY6_i7@bQB
z0JnhcWci|wHYj^YgPG2HC(RfnN14^hI1O~6!h<$Rk19PL6%2RR_TL`K-x5ms#dUJo
z*^Y+-_c&{t%5;UWOt)&4_7XQGPpVX>9a=?+mw6;r+|JKYIj|HWayK;kmLB-!w{NCP
zk*#u0>HG-#OxjmV((|E!n10;PP=TR`1XPI+Y0LIIQf_OHde0}^I
zrqbx%6s?I|TJN}gdy6yPwmpHwZMyO(F>uIWwC&__HiM7Obk(`dJuJiLkKe&DDQX#>5$B
zn2`v==S~!?t*u3s9PFYLx-Mqw*8>4J;}@gaT0bjPcMYPpKQZ=Sd40KQ6W5^dWSmO-
zbu8k|EzZ#ULDp;Rcd&BPoRdL@>!Ze97eLUEfso2#bmsMo5sxJ1EE6zfaOo2-^pTpx
z0m%xXE5yz<7RQJQd|PI`cV}Nh4I1nbyHB9ib@Hr}XvI}TW})eoF{mcYAXpT+*`W*u
z*4+0X2NZi$tNyn+Kb6_;3#B&L!k;!V8(70KSB7vhU5>T#+e?#T72i(n)*GP#v;D4f
z5$1CuE5n=Dpe}Ys6BFSdYbq(nQ08?#Ea;ySU
zDX;4&?*1wUtcwiJLt7cO^%mBguRMoRFAO`Au~|_+rcY-dscISWIDm$(>|avhR*lRN
zZ(um~X)kk(3f8k|9=e4_AD$SQbtpD&HGYg5F-Yw=HmSy&q~sf~rAH@Tg@q6Y8KQJo
zqpO_jHOU;)JqA5XHP;sl!$phSnv&JF8*{PJ$vY8MTLchm>gW_!$}QHZmcs;$PB)cDPumiU{VhVp$yJlGKpddqRe
z8QXc&J_D_yQ=)mALmTdFugZDaE#3xRm_d5+3)=y8?#*b2*=D*H+MsHttrIP2lRq0J
z#|t(f3V}b;h6qiKWon&x)-UcakA2Zl(duNkn*s_MJ@+20{VC9xj~(}+ihcdOw@HdG
z&B{XanQB6q7~q_5(8+OIMw{PPDT%iio2}Q+X?OB2Fg67pNfoEy7a8^S{*fYY_IFI_
z{HiX)f)8|{63y0CGlK5c3<5Z}Y|(P4=c~h|o}WbreVX39zxpW1m5F-zW-2JN>Yp)N
zWJml*ahkOT?I~z@-UhGvXaqp)+VAgV-91VdaW-`Z~4AgTJf7YFwLmB
zxNecBj`s4O(h%VuxAwvuvl_CG`uw1`vK((H*Eu?hecOd%_%>~t*0Kf)Va=A*xv!HQ
z5A#C74ZJxUqMXB88OA1Va0xQDFt|8xV#Zm3r(RvKN3i%hZ`qA`{QAClrS)Rx|`Ye0TT}~T*_lW1MVj(DBXJ(x0K>sJ)0Jt9a
z?vmT@k@-6aD`#FO8z-`EQiHHZy6<9caf#tDS&nUf+fvN&v+?$WceD>VI*DSgK0_QT
za-@GRYEJL)-hK3uv5E1N_Gq;({h(wWzUZlm94n!ZO7F_OJE8j(Hx=_z175e#NWQHm
zQ@|Q8!#5s&!gKx-`{CQn*JNeLRg|wTIzJ
zIoz(-%Wv*|x6>-j76lesI<5chQ5F@S6CK(BHXt;nhIs4RsC+%wjluXzdXDteYS{hU0`-
zKZWBVKf~f>+tCRaOIOizVZHhhN8=YD7`MZ@EcUw-Q?6xHcJ8zmhks}K+j
zx8~8fJI26bOQ1iKONbQh8OI15^$Y9wD057E<&oJhP-MB<`vzOuHZwCMWX~WUyI7y_
z*mIvO)eeR;*hYYrZugdG&9(8x>%@8{l`$HDwi_3gCV6`WDI&`Oah0Z(l
zTHAedvdeXz=D{JuY_hHx-s%32A+%V(OmfO6)5Ych?A*(wn=_4w!xqctH@_cwBHDDr
zc>TG6J~f?!8SALg#k(VL_3DAI^(BS98F;E*lznl3!3RnX?H%2o$!&I$rZN)lR*Xj|
zzIU1bDD*7+mbxNp4Zxof@7IqOB)Y}1co;DnoH9x#%3Clnu%bfH!l|{9*WQ%lbB_2h
zPpgQ{!ViBT1U$AJH^&YWiMv$sQ~nD8gLtM}jW;0l!@xk>IiF^dVzjO7*Ys1rdRuU5
zisg_gKS@4xX$jUx%U;Dg!L3Z$4X30~Plbji?21xm5hLhjn+
z{2sRyrcc|a-$^f7a6^KZY<`+OH%@vWi#QI{?YiV^&;iO{FKy+S^4!<)5sZJ))u!{3
zbO_?UXW+(BkDbulY)y~qCq#<7NS0IRH7Eu`jle2c3M#f3Lxb>8ftv&?d{aESiRS_#
zTDc*`DK0ZMKlP?udtIZo_a!t~YZ;*LI(7-0emnLYxpFsU)QWj#pO!cn)tafP%dzL^
zH)!`OzJdnkE7WX5zpmBB+fNEvDm?g8Jjn4{cBA5B>gzOFDI>EHm5?t>YhI&|*)qr4
zWcyDm1cs$L`BG`Ud84Jc%JyCuojv)4_|wXg+V9Xto~RQa300>VqFt<)G$pj~e0ezg
zm{qE|(z-gjy9H)n$W`wo)xXFM@{oTWXK_<@>u1p^eqQpD$`?!C(j}1zH^IW@+J#7#
z)18{kku)1c5@2{TLp)t6GPktgzWVLzrT%z+<-SMR@t*+O&N!abCvd+F72SgN7!4KE
zbXGh`>J=9ClDHq%57KFcR-Cx&8!cOu*Glim0?s04D#6VRRC;cqBtMl+p~+hyP6oKT
zWwH>&*RL6AI*3ee&oLreM&a5i?U~spHTeY9k_o)iy&5pm+^~v7s}Lf|KW!Nwod?Nt
zr>lvMZu?fc@pHp
zEt2P!qB6qxGFV}DDX(T0!Gf0Pd>iPQ@o_XE!jKj7y5edi3~ikC3Yoe!BfkQNP{iD)P9m?$n&@k+b7F*|8)ib
zC;9JIw3~1wfWXhycCS(^$}7tSy|PpE4XhgVSqMq5;BSd^%Tz0AZlMn2`@Cf+9kEod
z>|r<`R3*}mT*|ysBVj=)gk7t!3l0Dz-NG;|X5Vvoc2Jq_%4i
zTH&goPIO+U%N`<6h#BFZ4efoEEUfUs!LYq5>ROpXvql`8Y&PP
zqD0stf-mpd<-zE~u~fm9n~($WL7w#QZk^hO7b+9SGkaV%98W&BD8Cj;7Ru;!VRjQd
zzeM4$kM}?@+}EH!a?0KiRutX)jTUn%!xaPL;BS3TFP~FEYoDjzb^(IY4J)zd>c{Ua
z@WrVXE~tavZI0vZUxUbk-j-vGd*yZeK345{GKd_wJXmIabNeAdgr0Wgx55fRLH@g8
zZi01~UX-OdW)~%DhMJA$hNT$K@KTQmzomN0R4(M+7c=Xi67}9)=uud2MzUY?0~)QT
zpnux`vkqu~_{nrA^I_u3HHtBs-I=gxDTaI*aU;7K@#T@Eo4O*}QWi84rOUC_p@)Kb
zRa5L3PJQf*>m^|J5mz9fSl=n4_@X#u1a}XUwZ0ky10yTEa~#H>Pf6}W7*!TD7v6*)
zPN=Fh-YVHidcZjR7D>^+xCe)&LScx8c2bNd(ngr5|DrF;|Ej9Oue=M#C@e{;Izmd5
zF2D#0>Mayg+0hny@2`2BGz#f-l)MsmB@5zj&(b^g&Ty5wlujLKx?v0ZhyC3UGS4CcdTDX6Ay
z%kwyDomYx~yacO_s$`%+grPVh|X6yNybMPV78Q{k>U=jcYC?|VoV~QZ*I)~EHShOJJE$%4XI6<`blJ_!
zSSV4hAH|7a`6?g#Itp?S@qNDl;z~t~a3OAqjr@m;(`kh(1!htemZ{)a2Axr>;(D73
z%bD_zuHYs4E_7R6bfC8pW31UIKvY^)0Jq7D_Qc5aG`uV7t&oq0P$kU%DfO-Y8JPDF
z2UIR#$}yMn7-`M-V0Iu#gRTmtC$xGGNE(6pi&0gFD2&fh
zqNUodlCT^bbhGs>!REbeT$yxjFJ!B0VA1&Z!Yyr|`daR^NZiNY(Iq{0@!hfd1Lg?f
z2;#uBU2CAx_);{y#GHjrU!95Qkjj&^c
z^T)CLd6#hzM0kj~_CJ)%{{-z62%7OQl;)V6O>)t$6}pLUQc%e2jEDA?PB-_M36j4BH=SLB3c&
zd7J#AOt^lt8z1rNLtU6|1D8;NeR|n~pev48XVV<>3~uENetyY54_RlLKB}vd8C^xk
z9l}Cvq~kI!%|@HRzfW@AWWkHO=-eO53VtuyL*Ya&YnZrAQJUX+KbDK*i#HaS{i036
z(m6gnjj4Q6lm0~QP_diwmE-bX*5AA)j*)NOtf*ksK~H`)%G9)t{U`Od0sKcA4uI)2#p@*{(jC^JZjja$f-zLy8@*tuSH2cLC-uCPoG@>oBh@aW@
zIYY!V`{H%FcJ`f>^2!fhUV1=t
znlfZ2KSdl`c((tlHqBB)WBF!Laih%uLdST(IDG&0MuIQ*nl6T>RM}wg3NX-bMgyg9
zW5+x~>J=A|vCcXOlY?0g7XjH)ndBarq@&}PxjoD2TpyM^qdF;q2Gd>`TbpBc{Q3+y
z)sJQimWfh-rGB)Mn_hckEFap|O{3F4lVAN|WPLF^v-C%u$L(&SJ)?l&xZc65xejdU
z#MvX61LsY%FE&-`uFUuF_YB?U2kqbU-@STB8^ZdG0jDnuj<`2*_B7Q+vhR)Bnx56g
zo49!YD@UXK*XnpwiK^R%o-g2FDPdIIJ0;uR5xG3$SP7M3i7rs&$TDc
z9$U?vo1keYG3Hv*C%nu=Zd^#I5#vpwHnYj?`%5ftc0XJ43r>m9qL>z5s3mz^PvKC3
z>y{rl@!JWGL<+TUh7r6JT{6cQ?F1_r&z>?!-t*QlY%QzQ0}?iFAYtph>({AWS^VmE
zXQJ&UcwV$erQ2AWdro)i!l$YRY(GAa+edd`Xf4-+CCp(zq?noiOd|@L-xWvuTQ=l!
zhYuzGtdQ&$p&behQ7No)q|hkafgsgJ3H)k1+rgNs4Pk&T(-R^;?fou)yu#t+s(j68
zaq%H%`N-Pf3FGVgA`fpr?wC&*fpprnGhY>S0rbA0U;;X-;P~Us$E{g5BD*Jvnl3G2
z*q{CAz;BH0Ix<5PXs%_*Tk!jfgacn>-fLuJPmblo$*OvD`lP~T)zVVfRBoVxZuC8}
zcZ@~0(N@Ppk-WTjX**rnfB^OzW@2+AwIl*Z;&90iLNawYoc9iwz(#c?=JoFF{^
zg9eROgW`C7+GR(n!Jx_Wh&UMuFpu50_kCtS@vb{ZTUe$y7bxC|mJZ))ik~#BPMffe
zs=eoyERbB36meiFplWGyb@w0M%l{4cXBEjFdU_4|xFR}%6<&81*aTLk`Uko+=?<5q
z@9Nws;kvIz#(D(GYsEm+_XTS)3Mktic6?=?U&`{x)-=?9fMTD)(0TUW>b=?~NXkN&WL3pCKn9rv>+w$3%tH*{$4_!9C?kr*X`BWXbm
z>!K*5yiNJKE_~h>FD{s_8q6O?^l{q!pLwlMNYQs8ypT@)5IF~+C__Bn48&Elw
zuhV+;J%U`L%D^p)o>6`q(J*OH0Rb#3WZAsF;WH|*3xy@Y&{qJZHqXr}Iy2R`0V`~_
z$+r%$)k0C}4*QHJc)A{pn+9oH&va-%U)tjq0$lr@DIlS6yka;iJvQitU#9amI^_;e
z8B8x7al6-jhZnJ;?V93foaN8MgGyE^KyzG0R02Mg&tS@1*kZM(U213TO
z0PGQc*gaxxuP(Gzny5;AJXIp9F$Y-FYNCUO)|Whc&Fuwc4(R_a9|Y=I|H=oS{d+$6
zbLCzuU9;2C4}sl3P)4hTe!l4S5PpYMYvk2Dc`U~jAwW`L)f~xN6zwC@LqU6M8lITw
z59!CXaR1rE?#}}JWODU-b|we$l(WMAZ8O0#HTAo};AcQ8;V`66y~*Ln>TvJ4e+6o^
zcjVE}MLH{6F>|4LBp*~UcGL$r9Es`F8JRg5H~wN`?#oiO3%!1{f1;ag@i56^v5LfO
z5xq~XXljOxp655>kD(VfW#Xx8Wkko&vcjZy^Zi=__ppP+6r{1HhKu}Th+^SZ|G8c+A1P9JoHXUH(Sfk_g=$8HMvlXU-lC`Fk)
zA${W~y0YhmBkA&U8VaSReN(CBa{J7nHz3AHF63`}Nn>{21(~-ewYGoC6km;^qbyiX
z#@8ShfBpbjqr-)??F#+2WoE=JBl{ba8yhWtbr~*y?y8aKEE0Vg)%y8Ig7udV;{%B?
zrKd~_zn%)qsGqcM2d%3p#LnPSN`%{D+Qo9g4~^*1)xw=1@=AZ~Zy!6v8!N6Bs?}mh
zSDdKLO{Mk}+QrYdrXRDtSo{3STo{G8pN8IsPBKabT&%x|yLn%{Frk=Sr*UyR@|ACu
z@-@hOSaKO*qkGB7l)Bxo5J6Nn9sso{^XeE(mUzF|=2+CF7RS=zwA}CdT-*L4pxe)H
zMb?P+&nLawSkp%hBl>m1
z_x5M^0PNO=POkLNAa|nkg6hn-(XWGSOZlX+uDy=I5ytk>HQV}&`(c!Ajv5YD-lVfG
zo8wbtH~`sQrsya2QXZFk<)+}CFqYkJ6_bd^{_MBa52kID%vhj!B9V!A>5-&##Qobh
za%9{B&{%sQ^qJY^9x3l>^}kY>%6j`L=>6M^WI^p<3e$*3k@$<}C$BAGc3Xo3jH7&x
z^RZt1eE}Ayqe8!aEFI@Vz61FB`|N8_?_0-f
zP@(?xJAk@ug49Gqy4j;vL@!+cK+Du(RsE7ZH9_zokuhlWd>67<4YTV{pE~S7fY&P6
zDBDr02hrYw_HWdJ6u!+nq&rI7xBZx=@qi1&(r{Ti&P`svvnQ
zH+66D#9Q9W@b|XdJVQ7ZhRm5Uw_)g!_?N~#I3CpV`_SE}f_P_{mlIA;I+m5X2P*ir
zqpF*u(9KH-vco4Yew%+EKM=LPL>1AGNO;?)!#m3u~2@1JwknJUD@c<
z?r-Ku1z2c%o1I!a&miS
zSLO2lBJ6goA(qjY7WDUN$(BIF&PGj1D`yJzUkV{K3Jbk`vr`y_4Z_Z8sg@?wvm4i-
zz|8e5^^-zb6a4XK2d5Q<$^Hg;>2j6g>vKu{4&xfjOm@1Fn$a^7ea7rMGld4zx6`}M
zzHWGjYp|EpTbVgniAW^-mskgi8#nZ3wftrm+b
zhCS5}5f3VAn|ylu!$RfuF@)tB^bA_^e9>cBeSAkC5b=0eyWF3T?;IikfU?#fMY-z2
zH;K61H}dNnfxS%pcdaZhW>m9XL_lM&k0B=4Ac6XwqLsfiZPw-ROlKzZBLUI(>As3CISV;w)k(mk8s
zhY<^&+Z1O0sT@#eR*}GL0I)k_qf3^jJc2kIBq*))Xx;V0;~fyT<6ZnTO*)&2%fq8a
zh$60@K!&g$@9x@tqN<>cI8iS;1r0w$hFXnE3xn;k;&fmY#heKtr&y#c!zPS2dm-@-
zUN&1UHu^{OpHZ6(CtCd6lGxa3`#`abv)ntm%L$*Byz!+}FQqk4Wdy%pT!W1E;dP9u
zc)#bA+<{1S3M{X#m{nTyV5#-bpWk4k7zMjqYc`W+($Rl
zgg0ts#?lTYiV@I{Gww34p~x@H;sSWq#Jp9OMgHwBTYfGfPh1H5W5r9qk>f7RUC-xidd1f-+PltT0jkB(Fb&0P`78
z)>q7SkzZu@!nc%-UD7hz%1s!DV>aOyziWkm;+y0Ph(<=IucgP$UyGoyTioI`sn*^+hK7df1b5zcTH3xK*3i^kQr&DJj3@3GdmwNj~+eyRPgw(-PZlbmZSC$v@pyW;*|C?vmP_@^rT
zAFIN&&~@`W4ratC<%A7ky4?e#u_S+rd~45YtK*W7-F=q^v`*q>odhf%ys2aqs6(oVN--KfX
zWqK26F|git!HFQIT&782Jc%l!jyu?K)+Lu)gl^v$Q$vYl3w{beplZLQ=_Lh`l;0psh2^v9}Oe}r;FSWBr?OHL=&++fx!a&+?EG|
z_y#Bfqku$aU?}eZ9AxuR{bA^A>l7UU_jeQTB~{F9Hl}5xIQ?r*l~)|!gZv0T^!T4|
z`u!`i$qKv@Oa4^0hbo%CAzke+QM>&w=+&dQ`n}ut8kZS6w>AlOmBMw?eigRlj-n*c
z#gkrvjGKfXawu*Qy<5!NGbg~l!L)(nPSs5{xS1Q5dQx(Cz6W`us5g7)J)ad8I?b~r
z`?G3F?fUA6pH<6&uL2^b4~8bT&h#VfeZ09R6o*NwcYo+IbM2@dWnVz#6SSJTX<1d_
zrIM%h8PNf$=<^%zJFh`Z8nTM$6BesGiQw^iEvIoeRUc>grtwU}j%?pe3W*Dg+=k}6
zS?sMs?)Cns2yK;q_rzSWcjr@gWSvyx4uSpC%2O#C+dU6LhO~_FR++a#m54o>`;mG
zi8?b&@!S@8X-T&e!L};M`v?eKsi4KfHB*(@&T>~0x((;$`ii!Uz_49;JGLYEajT(!
za9_Ww6WqHR*gdBCvrJaQOI1P00IqsI<9zy4=nI=z-XyKWfR6+<1*EI?n*{fO!k?=b
zK(NGlRpHdl);PX9swtSU23~QIKf!r8ka{mSMLylUp!}gD!n=GI|I04+-pq`(7kBcj
z>QIWC60g%9feygDIP>()Wd~yRoRRA(~e!<5Si>A1B1t9TD;}vW+P^}64Ui%D17O!hbzDe@MToGc;Y-&
zGi*%K28n=z&;4JHgWi__%}JoXWI63a=p15pma|&%=MKdZ{>9XVQ;|p0;20!F$=Qo3
zYtmFgCH&>Hge|=Oy%!PZi7mDXZHRF`aH_4k82Me6s?XUW_!qCWOk7NC!Mv-(;424K
zgbnMQX7vj8Z+$1!;MC&GeU2=by%@a(xeYZkd-hunVSWM
zc-6IKV8>q}6XKE=E{Ls}KLjDh@HUCSp|q!G9oL}v8L>fE!lfZbZj#Ort-3IYMCH`;
zqvGEzp)Y6mOwA*(hTi!r#GF1v-xS;ziigggmi)}CQ)WjMr9mAD(qMuD_t)YzpKpSn
zAt8&FX#U!(Z*@7$I+Nak5DLt&N`xP@inmTb8qEh(_TAP1X2M-U6S`a68Bud^c10}h
z)fcFjqZax6C)y-KpJdQ};>+NDglOb7h*l^_>n&MB`m>^T-6fDhyzKsAfX>9$ko4Sn
z6=3(Ef5tzA(5I}~_A8m=ZbMiwFHr|VSv_PwJ9%=uh3w9p5Ddp1){hLi>@w$`m`fe9
zH&h+ySYKM$htQ~^O#45avQxfr`>`aPyz*E`r_$ny-C1Af`^Qvd>>wgS^}AK{m$taR
z2J`xz)WYA=^1bj#MeF=-NxicPBVrDCT&EQ*r{h166clUG!dG>5XbTFXof|`A&3?T0J+U4xc
zIFM)nanE}>>$Mo8p|QWO^nJC5?1ekQ%!au*Z~SBY{}C%$_P9Vp?%|eJ+3KMhFMh&-
zcU3jOn$dyIN?Z*vSlGR38af|h7PT0bw}n9|-|1{2S!g2k)K!c##7S79D`{|}3CmxK
z32=;vSa(W0WVFb!q!v_dWA^!FA<&7uiF}<*g7d~~0;_`>s7QZ;v(&+FV*+1K1idJ6
zY{8=0W2L2iei}tFB
zyJJ4FbKi7s>0PsQ$K5buvdSVuQ*}K(J?)1CYx=b&i$U5~O*kb-qqMBA0I;vV>Fnx#
zfL;htb#`?n(E55$=zv=GC*8enwR65VLV!kPvi(jv0($5Ah(CgNV3Rd
zku3{+LH=%j5PnS61n`QYvv{o1S3mfMW_Ufh-paW7Z2si##DO)B5z0AyK3hd?!I18?
zbJLQ(C+LwJFq9_ZPA9lpxCH@|;*%}RrsvC90QNda5eVlCxN4pef(!W!RtF5Iz1c;}D
zCy8QstNeU}zwh`dYT$wvFv7
z+b|EuY!QmT0jB1F^_<1Yhm179a`_q?r|OsSweucjy>8CXkUH*5(qQ@KlaRr^$0%f&
z(PNrXn|YB~W-}-UE;37R2Fba;)jf1cChXVwac}=+Q{%A;jrIPxqWR`BxZ45Y?^PXN
z=?e-(%Ao;w6ty~CpTg9X`RL=M1pp7tjsq4`S=laxKHea|Ic`_?a&Z`Mky}3!_k+EG
zpnTH4SplT`ozZGcX}m4Sb;F$bcgWpjLzi!i(@lN$%P!F;O1_(_2v+Y6$B%Dold!Fu
zp3<7n@e?$G?eIHL;i=dO>6Z@?dxpoOyJCDPlkk8lq4mNbH@dWj=6g2z{&iF>lx0ATIAJV7~39f)7`P
zXTq0CLGRW7#@>6zHP!9wq96#02o|c8s30gsnsg*0O0j_S-aAMyQWF6IMIeA8pg=&R
zcPXJMC?)hDT{?u`y9p%m41C|U_ged`bMOA$k9VIBd?U%soHKJg<7t1fqiL9zLx8YN
zw9N3rg;ES8Z;`-^<;&I`FQ}wkKaEg_zW$)W3gx
z|0qAiO_+YcA({uR!;6DscC$@yAb6$OB6fKIw{hhi|Is3ETq7h0H?ZA3KOq`ALJCQC
z`T8JYi48eVV!$&7_WkTEH-}k*V-+!kB*hVoNjZmc8u>})31gcmXh(K*@@rZTspZZ0
z1@1TO$1ldf=?$L^E!|tbVqj7mW(pPNqqzJ5CdK;ELHINIz&2i+DQ7dp&69Wos36=1
zq9&ezA2ZBv`=lbDOvtqz6Ns3CbYxkJ%2?&7_khnlStWnJlsYtuWDb9%BlgZ@a!%*<
zvRLq#6bac^uyN+n3IE;0Ibfw|JsryRZn=MlB}Y}*q$6DjRzJYYw>j2;Cp!FMwV}H$
z3@3kRWbiWV!<>NM?R=5&Sn!B69*DOJ;a0U?Bs~9P!e$C^qw3D(;y-HBg3b{4lj8{Nn5pI(Y+d%kOJGmDKIZ^5(JsYsNnRg
zkHbp40&>v{4&{<<)cDzWNG)jrXpyXfO>t~iIc9B^nH90y7+NVp*hR^Wkwc-~g@W%#
zKAx@cEP##A-mo5>i>ZTW!!z>I-*hIhc-x@Cvurru`LYAG6w7b2Rjp7Li!l`7=7xZu
zCpgQaC814vX(|bg({>V$&cx`F(kTdec_QXNS@CXgnmoCP9z+TI@okqbG5Ga>zAJ(Y
z`4l{g49CmI*u5$ENpN5Roq6LkD|O8i4@x9-WYs=+kONuhG{GOvck4@%evB%sfr*;*
zB4e4A=Rwl2obWVYIy>K8?on14UXH`8-hUBIJH90YX^ge+`*p|L$MDm;6}s6|(p<}Z
zU-%p0td&rjw(~YTC!ITEGRgGf3Zj!SXLlC&srR#+0(-j{@zSK~Pn%93E1{3%&%y?E
zVOOgwP_;HMrJ4$(2SSuQH}^Il9$ApuFG;QkWh{*rBZ^1uIfZ{%yqEQxtI~_y;V}Vb
z$|#lm&ty8m=4sZ=N-a3TcAxjy7kQP%%H-Hq4~ev-l3e*yq#yC>E3bJd=0R
zggr#|vI1m9mbr?6tnVDnKaGf3Cs74@Re#+=ybhG~IrL5?<+0t)McEi$_;BkZG;%|ZHx~y3NW;5~IkuBwl_l=d?d?+Y-eyJW
zMz(KSv3dYyIMM%HR3;X9O|WCi${sTB6L)Az1{bjiI`O;+>7$
zDcw16b6wUr;Om57dhLzfkM{Gi3DQr1;z8s(0*hHFokRnmiJwL)&$SG6q(s9*ShDSJ)DoH4VjBh8Cd_GwkH+JV!$z0vAl;hs5MJ=Unb*62HzpWcw!zP=aiu
zjTb(st}*Qo4|~!Sx0sKMXqG8{GZQ4M(C@r{bF=X2FGY**z
zgm;Os+YSRmcLenOaczzni#UeK7tw4;A0`+mYiE};PK<9=I=V%N5~49us1se#PDtq#
z=IXwtN|Os!fFJ9OzsZ2m7%21)p^-zcqqbpSQ}B&iMR{dajIO43+mR2xT&%6niX5gh
z`cPKZp8N;`Lyror$^G2*gV~_uB^C^PQcqlJaIq1YxHwP{5QtFM+{}!h;GYE=g)L;1k*DAgjyg+Y&#F`
z_w7fkYYwgx{vX9g;^^D?RcEq(Vt9j9b%Qm4+f1n1qOWzcss8pi#raqDKSi<(Nj-mP
zO}pri<{B4Hv=n>gr%^$^cY7g&14TPx)oA$mp~gd%c1yA4lmL$)X`)lB{Ea}p(yd~o
z+^`s>MeBmO*Z>(-kVC;}*SQ8Jd2vIW{CI0$K9DG`U*2t(=0fhqnWZ5)zRDTnkXi|X
z3qPsWum#WNkf!R(PYBkydCyrB!gORADl@^hv+A
zK46))FaDwD*sgeDu8GH;rTT$i@r71o({#`Woqm$4;Va!KL=n<0u=Jnmp8KvUC52hk
zHdUM}hx`5}yAlb@k{VZnw=rriSB>TL;}_CEOx?hGm-X9X8o#<)gpdCIGJczQ*0$y1
zbyI08q|2S6>c;51p%}9^b`Z7*X59BfSyKc@!4i-;C72t_yYRb?PT3b8DJ-euv?T;l
zJ-1)O*FFNH)6Gx4owd_MRiZNP$K|sgp2O0d^~W%#sQ!b?DNG(wp@4cm&JzewC6Pee
z`+0WYL$Mn@D|#=_M(Y+gC{9YN{*`M3T^6`qPRQ4vBmQJXOOud809sNMv!H2ci9ZmF
zPb$KytlrcyJZM!>gr|jkG{(R1nIJ)DKw#v=3Z_c=i7Z^5^hKlg&9BfO2=(uu?8a`$
zPYp+whzO9T=U@f^ZBfVFisI1UXU)Fj))im}=?HWF(FuGleAa5(Z^pbYjdm#lhRW>n
zYWHKuZuI^EPwK5}*=$OxhB4!o&N+w=&{UAmKYf#BOC*j2Z~f@HuUG>i)u!ptX4Hj~
zyvfi>dW8KyhXc`=3P@Z1_G^=Lt4!GX@?OskX8!7fInCBAZD3R$zRP9*OGV#9?3ijy
z5OBQmqHJZL>Ic))^Jp9}*@Qj%f$x`MzrNW22U+_u)O=%E-+b(v93n))N>~|q&d)hV
zd#Eka`XGLs{YO#pbJus=NpzhUoc1~-kyYzC{cShDtZ20%m*mC1{4UrLX9WONT=Cwl
zfYwazH2?_Bd#K?y@`hs@F%Pa7#a&i;jWjsWCrP!D3h&M$-@0$E
zYKW~z(`)s8ihHd>hJ|85r2^Y6!MXWYYe}n1*DC^VkO2-P4rfQ;WAm(hDALG}+iT37
zwlxr@Zvah}m&7Z7C>6%LL>mVIMkNNGkzH_VL63c7juhC>6`!-&qDMUDHs$xaI~`3%U$V6FCWIra!+m~zR^
zF_Ps$`i_+$;~JW%stUzVKi!~Fnxmap;i-AK%c+W`rVK!m8g`4UT}4GOJvjt$YXM#p
zW4|TTx^ATSeuBH&L6ey_AIB9!C!a#IHx*@eh9G%l;j5=qUo|X9kj6gdlg_NFem}0t
zIR5m^8NX!Zr2Jxr0-LJS*N9_GEO^yo=(7ci2x{s*=uj=HQ7CjkMF}KhvZ@
zcG3Rfsl^dN;__w0mkrMA5P*QgpOR<@_Fw9^8eZLE%PuV&@V-9!P`zBN)jZq?A#$)T
z`gM4;^b}WAMC+L#wj(=Ot1c*uNS?Wu*rLaIaUc*X-lQn=M_F`-Bt>)LMkJOjuY^vM
zP3Mi%kYI#+TkMdEp*NrWYF`*aU)^f^6f=6D!kAeT9X=na3z5>eOeri_s><8eQZNC^
z6(M8K?&B0^K0rgJF1c*V3C)_Em@gFY7JMKPU{)@sJf|o?vRWToI6fkQ^XT2_6KXHS
ztrknzeIc`EgbN7lyM$RGIuQ7HwYMJdxVoa^>(}zJvwMo|PA_>-7w~bnJXqL8Rm0ij
zt&+JouK_jk;XkX9ynm~a?LdvJ_%~|gR4JqizNUA~O3DLP(wM~X*SB}8y1VlaN{$75
z`@bl;KOhnFH+Yp=KBfxu{tH$3??K}JKZ8V_{{|uVdBTvO@vk89pZU0d2NH4rmyi2z
zb}9QWHcf^8CSwMho+%cmXQq(SLDFTXgDtdtPTq=4$h$Ma?01bmB$El~NaQn*>q<0!
zXZPcPfnDdfi~CO3z13fjI!zj7XS&lRc%J?6;G@IU=;9$&MzQ!@o0m)!WNpu}Zs}*u
zMwGESiEGO4C4qFfs?mV>_5E(B&R2Kw=*H@pQ*u@B1luBMrFT(D)yXEA_NAln
z%3tg4nWP_+^slE&)`&bsv%kq0aIN;vH-!Y&vMUg4pW6BxOZe}tG%KZ9cnRY#Jgli*
z2fiB+D*}G9@QU8O$68IOgG;#FbZ
zRkzO4KQa#h1qo}d7NER(0u}C+oc+$*!843zOilSNoj(m>9=E#v-s}R~Yny5Z#Yc@i
zEGIR3RDW?FteYgRGj1Q4JUDcEC_V69^5j^xaF4L+Aw{oGNkS!mo)C}UHG%G5F>3yE
zKfD7z#wi_6m%6~VdjfX+#A82G*||R5-EUFnUNGRXXrrLmViiyO>>M2<)pNY;710dW
zvLu$TTcyLT&Yq2MU;fMR_-BqH^Sl&Z-!Ybr3
zTT;qFt9*`lkkz2Z*w
z^3>k9V@#niP#pa_#51_=aIpQug4Pr($@_k=z~c7)a%5^s#RFwoKWL17-+{S|(u9OR
z9dan{1FiCrc64fPQ_gQPeT}QL3J9>(zkSS&Jft&ZVxMzAq=~>ie6~thK3H@*=PyaU
z51Dy1Z=DbT7|)4Qs-6qvW2}S-z@?>Y)mc!dU_sbob?}u;_(8XB+=Fs#_F6EAA{hkB
zK_%8aPNC-??X=l)AfUMnN0Pd+c`VY61b?k!E|EHjI;m1n$crK?HcSuV>rV94V5N*P
z2|wNB-nark9JFL1w6)0pNTI{8mJJ_uK?Va|rd
zEPa%`clB)!`sJv81OE~m*M+k{+2gda*Xs)Vwu+nwXe%q(=VLLDg!i>F$*h4$
zPsOU8_j43n@!spG>nmtzBA#LIOE2Sg&GSvf>ifX)QVtZ{jPymGVL}8wp@HbgG=DU*
z+%4#1#r0u64AZd3c#36gAG3(r%e&by7~IcG;1fSt%ud${0N8oUlbh>=mD4P$UP`{p
zM8_YTkL%L#==oqB=WhqyMFrk+%FN(Sa5YOrODeqI)pzcO6@N@K;wyy
z6SY{MAZRXh2Gna&ed<9wa8yHv3KVz_Fz{_4kb2L3uxMQIBj6a%zTZf6m?kH|j{)>q
zUCUCe)Y&!yz`sMe;#_r|XlhoHQbeq)M#?%{gFL+GzF!h{v4D<#o(QCE^KR#x8iExs
z>D0zyp@#!io-U6hWtG%RjNxHgTT5Kr+clntkrxy>`me8|r~wRI)vW!}d&IN>eI{Zy
zV^Kj=mM*B;VZrcHNbUd1mkDki+GsJ_Ww}-UNJCU6z6jB%H8f-;o~}7*iL4!wIt(S=
z$+h^fwtLHZhmLyMzyCe9V`mD=&u+B?#XfO-#5BX=W&~iJ!ZCua-*S!S?JLeya*Z>R
zg=OXvoCZk(1R3B|b`+#nd_Sxkad%tk14EgjT<>YFYF*9mti~4KBj$uB$^(&HOeHDr
zY8ul0G
zr3*MOmRo;wUhy;pckEQCjPiQwOpss&(RS?4cG_!(AEVv3$gy8uMt4#c3a>ETIc2VY
zIo>SpwfhpJjCZ{fxQe#cQH`51l6sTpx6LA4rqn(Ui0db>->>l{yyQW40Tc7^7|$Te
z$afaM-FeA48j%wCY+C&la_fOmjG^93x!d{6^1S_(>k_kleTRpbr;4G|&Q$UGbRe!v!Q-J-|XblBB^SI`h
zWv!Q#+hxgZpWz!U;LhqJ#hnMOeT3`rzQypa)e{4pFx`bb4
zktds^iB~Y|L2%%XW5pL~{&1HI
zBsoA~ZJ}E)m4wx6Wf|MQCnbu6Y_!gu{ZU0uc}~T~hOcAg!RODP<4;{95B?w6AV&3{
z6E70~ns_nvf1Y?Th%4XnY&APUu;2|NcxM><1{kSW{Vk(TC&Wv0*{U*MN%RmQPk1|%
z@LZmbV6aWy*zn4qKBiviTD|`yLgd|f?^Ews8RIhJB~9y0CB0--%vlon8qe?RJ(p)E
zfOR^glwdwAr#CBf-d>)ea%(*`H4O7pz}mH3;p#Z>bWyE!CTr??3N#{aC2;ellWiUc
z75{iS
zto$a^uCIE0+=)FZZJ?^f1-4*a7%tVQlXyJ`!8v#(dYwFzG$PLi1s~d2unK3xC7KSz
z-}ZcBIhC)bN^FLCbB|O^|dK`a9{a>Sy^+N%x>({XwpNb3cyfjXkO@x>dETdd)HeGA2>G
zJuATs@v5!D{L`3GBWz{mT)+OvyQsXe1^n~9%k1XEA4xyp6}rfoF{4g~F9sX>=0@2H
z7~XhUYdT^#ED-m>ta`|0p{Cl$Htw}~8kL$ac=kW=f7&4?UZk=A5YO!+h=JvB51D50
zCb}$23Ak^P{QEPBlMjR7k3fQ|$YzDvW^U(I;{*|4J-&5=sFI7N(dDhPQ*Am1##c{b
zYBrakS*qR=6X|aLB-`A{nhuon*WYC7%j%%O%i^Z`CIV#NlDey<7BvA~d-YLjZ)!aJjrQgT&(h@{KClaz*s$txKYZU)JFx&|mBuV8C6=GJ
zdu*9rw(*k??zk(zYK-N+2{qdiPqm)6KxbMO7d~%o;wV1^lB*X7O-Y5Z3R~<5G*#R~
z4A4;Z&!#9#izHPuvcFN8}wbcVEr?obsGIqLWrh=r;P=0^Dcv{|g
zdvz$kz)#ECs|WrBAWo~BLfaX0wh>RW)pY<;NICocj+9FSAd>RmulA-zgc8pJ-i@k#
zCxI)te?LcMESYuTY|9YdBeX%@_}
zxS0b=WL~S{i)q+rcA)A&4y=i4a??9P6Rv+2WTDHvW-Ubshid3mR?|#y9e4}emZ60K
zGhMov#)ASooDpgWOtr#ypQm7%BPqU*WSneKp)5B)Sh?@)&`VX7y8AqnlMwp{h)Rkr
znhZ1FlyMn_(vlPTe1-Gt=jfA2#`K*Gkpcx2VIrTk;(s1YL?Ha_&9TNBya;hw8
z9c@!->OIJS4$GKlxwrHfA!1)#BIB|X`gS6s6A-V+o`TV0J#K5eXGF2>s8Js%*LX$ak
z3^BY9P0+LNoM5MuUmL#9+^)6Ti$-JnJCeIqaHTc>j;R@=FN9-Y*2;v$=hXN(Ab
z@J>zK$TntAJd^j_1LDuJu~G{BLcS>8QED+~2?}abH=biYEV#!!Uj!)6&2^Gggy={r
zsN;qQQ0!iY^Y!U44NdRy)qTOEaiFs9?sYRs9cr>m)?A%0%8sv2xA@y`VbjHu`C-DL
zPWX$CZw&5J%faB7qihPVg=4>`L5&ByCyi}hHm>3{Mc1D^->Sdl+)8n>M@`tlsrL#t#$Kt6@NO{M_5m$Slr!bu+ql0n3L*{t1LHFp*l
zbO{ZIt@-L#oT0=%Xxtx`wAWi)U_hJ
zh)x2!Rmtba1ulZvml2JW#EVBxuofS)s+);3_u&7OiuifvAE^l3n4A6I+$A?7Zg-aV
zTLE-T{k!4)0mCmA$h)HCgJhyEZ;WDOv40k8i@0Z_Fnq+m6L;pkKc4faOPj;=qv@5(YZ~|L$f&dnS
zkS7F>zIuh3bv@N>`P32z6w|xJ;h_MFhLr@}HI$-{jYqioG*avO2IE+#FsbbeF75Xz}ue8O+#_B-tZ5HUg{iwjs+o_pCeC#e9
ztuDo2R2_L3Hapt9v0kZ%e1NAv$R=h9*0W_{ob7}qjoeXgn~{4Zqs&;woREq$h^hRC
zC>G3ZS^CxnUYKy;ox1Iq8c%nD!tcrgVk2H;u)75hCumD%={ZbsO$PBReysDh+~p#?
zX_>pZH$Oee{Zb2Q)TIi^cvj~ovc9TU%nK~`W1u-ORmj%xYC|nw2_k2Q@%d-g;?ukt
zq;#FS)mSUJZG$IeXRrLi0C(1K4iFIEt^Ru914U|RsC-v>dojfg&KIvrD<$}CnXnSo
z^~{But7df_4RTwg!~`yl-()H
zBWvI_da|=MB~FDJ`JS3_Syj%8K)yRuG9DhA)$
zsJZ>EvCw2ls$spE?86uP@nZ9f?0XdW!32}4_MOdZpY;Y+sHvB`@4G~@H={3qox)JU
zg#ae-ViLV-cJrPKpzF9Ir&O8^A7%O191x(%gaUSS=~MKT-(;9o!;rf}wjyn9x1m25
z0F1aAr>mX(@|zXH<+;`jr?I04d5pMK49v0_v2des(rO~ctt2LGT3hm>pU*DJq>G%;
z4r$YP=9IA~>cV}&51nTf47B<%L)3F(vdYTTb;2EJo6g-&etdqn)pv^{hsbGKS^dTq
zYL<86>*i}i5S*u_iEnG`i
zK*U|F`Y4WWJcb%l?8sJVVy4KvMQ60TNH_vL)9GRxf)83_(@x_*K$A({ac(M6Wz%63
zBuHN^|J*K<9mUBpOOh`XvsZ{)BMrLcMct1oTYv5eJfp#Phq@EqR!?7-uLrv%gG|jg
zEEWNuVUXt?+42|4mVJFfzxiccA||
zsv~q4#d^2}O)!gl$6zisO2!d^TH?zpG!2ZFOCBwX8|L0;KcfG$Ib@e`>GJ%3qbDHX
zy>%qT`6$HjSFsJJyR!s0?x*af;b}#S@JO{h#H$clWDCG6S(L)MNqh{ru?eCgR
z{+DWGEeF`Ak?J{+j`Gf%(<^FbCP4hv{zaNW^S$fHFfNXxhSWz{);g9`IR>+@T_Qt!u5S?H_
zK;ZhphWEbF3z0UJX(lO>QjxM-TWPhF)R{;18kI(@Y%hYePv_Vs&lW_Oos
z1UEQZaot)48}n~mKsph
z$L-Pj@zX_S0SEM)K}KT@K`dad_eg<`Kf~45ahhFd74w>1HtFXiYj0&L
zzX57%F{;9zCvtQrKqsDU>F2|1Uz&=!bqwzCO_xTgitLOua5^6e!G?QD?J)WZoZyj4
zOMx$cb1kLd&BWaCF;%#;dv=YYE6aQ8J~Lts1vc&dllOv!rO83C2bo_`>$&1`G2ma
zHn3$Xefcq3he-;1f3=e9oKn&Bkmq8w?&r!kQHY8Umg$3olTlCcnoFP{5)F3^k^OR)H8;v3SWyG7ODPWxv0
zeOhYnb<1jzj;l3QvCqGmIb9C=W^EiL)#3B?xgQ6Eh6-qoN&g@5*YE^j$5)VD12FXb
z3@E;uKX|k(%k1vo`ZdMNC&*V6lX<3pD`m|{aTzuIxPuC
zmp}MTX3Zh{=!h9~eXO>F1=UWZ#=-%(W7U{z_*CZf<1!`nX$Uq4s3WGISkxkb7J4!Jf>V`
z;T3uQ^c?v;GlmyRm5zKuig7bBX7yJ56-LJkBGyS3R>;=+lf0cYs8n`Qhq{4w^WN;9$4Z{l>>u*(`BeYo8qA{a{c!-1#BiPctF7$bk+3b$I#A|rcb*9$E7xgc=R
z7P)xx?^4ryU=e=1hj;UjdE4o;%aF<0faF+iWEtxNshJZd`E!t3xQOPx!fm|sFKjy}
zejThdhU!9HPy9L4rWOef+vQxeDWbcZ7_;xdKuI1K-Iy_BgzO}|=bB)&qfR;_ja|rnOa+IT^n2shjr$0skB#T2z
ze&$c!bzV4H?Lc;&Cp7RGWT>=Z&#R_P`(WetFv}Wq-`AQpTbom!HBul-0F&!9L2mI(
zKuQ+%y*c
zv+&h1uEXODN-|f5N7>!xxft-9Y`hdc@!f<+&;97@Pp#pdy!!a7ufE~00ZI$8!uJka
z8n=`3juZ|og=a3BTtCvV`d;SR4;>h+*5eD{n6}H&B9RR+zh2n%hFtGLrjbL^99694
z4HmDDkc408;
zyF^fKCaR=-J|6#sjtsp7ZLP{q4n$v$VwH$`rb9_aopG6DSzfWlSI+xvrl?
zniwtHjgi?1rvg4R(SG~z2dFI09)(w3t-W+H$Pqy!%e&ZMB
zMEROl<>`~kPDcrT#^xnb%NxFg21Bn$7M1U2YW(P!8T@G@{}O=)FU`;bp_ukc}Os0dBsS=Wn9o9n-RS`r?$)6oSi0
z66$y;C<9MI;HDsznujlVf9b2a-Br@Bp*>AF|CrPjFY~gF&J<^p*2H)4ip~1i&WOEqIeYsu)!E^(1DW?
z^);zIL96mvkgQjJX)denG+pRn|A$}6(%<(spVWt_G1$pz{*X=Iy`X%v8Aw#CgCA5{
zm3%cw?iz|@B%W`WJED$rH=>?xMuCV|W2J1Iee=F08Nr=+(r$v#i~aE#ehK9@wemI%
zwhMod>7zD?R(eWGBE7Br8G5b!s+6@w`|_b~cL6*dRd5bkWT|xKW;44TRn^rl_vR$|
zoIOT>_%iR$Dur>;ov}^>N)OSoXpWesSpx~!$~GkQ`4zg!3c=p~w43mpVh{2MElA1N
z3LxX-qE=CGP5#=g$%;dc_4=g|>!*k&s_G>yarzCrBrdX7G(nr9*z<5;IY0!DEci|K
zd9mfS850gRI^iNQo`2oOYJC|$2nC7ooBePnIL&hZi1CPT0NP(|#;$POupa3UYW>2X
zqpk5}S78+Q*cAkQCCmrYh@>J3IGWM#*YS_}{k&?xAi9WBG|O+GfpHo7E-FUOvg#o1
zd6A~S$x;P{(rXMr}f^;ir6GuGUU`Et^H2?>=LiouWT7)O!WssBe|-XWAkbANlx3_7qS3kInu2=L=gt=`U13O+@PSUg+L)zfs
z7f>tYzRkFf2nMkT`GE?Dy)Vj5wR?Ifo>ebT0n~XG&WRZtDv9VqIuFlek;(_V5#R+z
zIM)NFScyFvVSpNvJFRPlO>c
zpG0HmtJgI8B|NTB-|3uVuS704)9%w4dqYVtU2D9ZE$#W5JJoEcn~7`gM>j!)or@qe
zK)5XT7{m%Sz7Aa*{Q8rKJh=T~Iz&jw?gNkeoJ-H<>~gD97&t8JK$mS-w9wD0hnuX9
zkkP~Jh{DGo@8`CiuC57@4e--gd`j9;SWK8247u^NYQZ(%`%JK_%ZA!1RxmpFjvJgw
z90wjh^rj1spDQUylkL#s^ybvj5RmY9^o=>xiAxjXBJ1)@^B*3ES1PyoUj9*X3wP69
z_G^D-sXK!71?KMCF;kos_u4MSu83~9oPtxjWumTGePYRp?ukY19{1vYUXkt2+|l?x
z$uhe-+9p|l3ywLS3l+^=oFrat6%`u1I`nyC&Mo6xJIFA6;i4DF-uOxMB|1qDJ0{p+
zOplo4ajIAz&86eaBRS&y5|xjgs^3htq67rEGhY8TG!r$-48UwZWm
z?g(}YjA3BvOVZnNOZ~T@08l2~0?lOS`2(w6luI7ySW!MdxP2Ju(m%zLxgdSg0~Dm+
zt;WBlPgYa(Ja&U6zrJoWhfu4ulw&lGg6wvSIhO6+mPmsFZ-4LFo#v?0!~JicVxCG#
zZxS6fKT5Pq-MZxzqT1+s=hQM#xfvr0PR)7qNuY8#z@t_Bb(OTdMKxRz(44F3#rv94
zT)`t+nbAEO5ggCF|3YP)nXoQa)IH^5db7rbPdnUkJ_+AQ7kh(lER(ws29pGT#!&Tr
zk=fk?PR5uPdnG&r|Q&{deeDF)Z0uqU4&lz6RA{BVAOdWqT~Uf
zB|FZ)p>2uaC|kwk5ghxkC+psa$Pj0y@)nmvH7bQHZX3{J}rv;_5IZNlx=wWm1O+1#6AwDY87mjuYSgX2b{#SalD-
zR~2K&-0m|3s(1Y0}p8(B_{51cv2;UAU+#6tuW0*yikwv
z&V_1XetHqv1U;_&Ik&!4-L`tDlkn20m0HR5H<_u}>Ry4a&#Q_~e#>ifX&wQ4$G&21
z>!6Q!blm7C_7`)YGY>)Ig0(OXSSkI(L(S7m)OKS$R9{;TBf8lsfCp`>3br~bodvIB76M~jNf
zJ_k(+I}WQ(>DP9{$NkLh(z3HF@2sb|5A48iR?@m}ShT9EYv^*{pjGFb{^s#Denr+iRkX*wkGb!3sf$jz{Sm@
zZL+utb4e#bN%`+Ff$_1hMa=>4+Vj0*ZNibhZ6rnTt#qwz?4(HN$O$cWMiL7S^y3=A
zBuIh`e>kY)r2YA8KgSzjM7CI!-e!GvXa6JqgFCn)!PG;C`0f`jHH!$>j;|tdFMMvQ
zwyHy7i&tkLud;ggSB+l0aZqJIv?e5MyxxaChB_dB_Ms0NaJQXi*pTOpeLakV?sDj2
zo`g@hFxuXW7yW$C;8|Xb0X5wxiT-$UP0Wbmfz1#o;SE8VGH%nASQM;E9yT8qu4kV<
zThE3KXN>0fUQQ_GXS!ZrW6V+aArr@nJw;HM>D0a8I)>v6J$%0d$MMm(OkE^gBPN?y
zZ>8n9XFAICfDZ5WJOgzBX5DQ+PvdvtDiaUH#6M^7Ue2sWOVjHXM5A#A}_r~T`Gy%
zkM@6jb5ZlF#jQ!r`)8X<;}W|j1W?yLb;d4Sf5mE?0wxddPR7vPon6
zXK&V8#z`Kh*N5HA-b;}%-c6Dnws?ifW)(=nt^WZN@&3y+JO96@*%>VTnPzt)-}sZG
z%9EZFY`>+mO3y~_xD4J4Aq9r1B14aEfLLzS*zjP
z{5DDodnGCR;bIl}ew&BG=}v}a)zv<{S6*?zQ{DVuo^d;>*2V4uLlemSzLxrFZNG3E
z$RMkO1p#FXr>0{4jNKf>`Nh}^-FG+@Ez1oRzi}wfJTx*n=h}5f*anNpS=BCPOiGDD
z
z<3c;XvQ@?j`KMMl?+0W*l>)Px&1!o$S$mmS9~PSPJBB*xYXQTn4gQ*o&RO
z!OL-VEIn#)J|_BUJ2RKlx9vwJw0Dm~$!hjp&xP4bEC{=5hZ845#A!}Vpit`miJdf!
zE(;%ZIXFC^hGjILLRqH!0DjlsQM|dJ8Kgo&!P77}hz(CU*_d5d=$XgvwHWA|rCVGj
zl0kcpbGUC57p_lm&aBjvl*J9;?pLI*skXA%eQ_eDxXeh9V~mlFTD%c_&pc!ZO~~E7
zV*$TFl%GQMEMhQ%lKmIs<2W9ZhkQn)2#Z<({F*3a;hh52SY1u)(COm%c_p(vAF8x!bX5!((@@l
zbF$g4W2#Qzi?X^UoD(RC;XE1xRBh0^^m_$E#pF$A7jtf-GZpL(xo%xWhR(m)=c*2dy#HE_lw_WLK7WQL^@E
zl-J@E{=J@0E#E}j%~>kn=W`PMnCb<<$!bz@&Cy5gEjekfTs-i#T}h>OdG3DLU%Y1G
ztmu68_E@OTB0H?2c-{7<7Txl_cr|}2?ig&PGOZ0b)$lIFR+0Ww*g(%YbN1l=+b8c0
zeUv_T64OOpau=GH^UNu|>@I&$ANvk7=VS!l7U`DtWJ4~DV)q*~uSkE5WSNAaJlI3;
z{gzyjxOa%wIs0hb%P;$wJ!?$2$xJy
z2>1uVX#!yDIQt#}7QXO6`KZQ;fnb#s|?S4J>Ico$(
z$wH9-5aM0Y4_{1rJ9;lSJEBY48=Kvyym}veTuZ#V_*WgQ^!_bOS-Z$l-Nh0)yRD*c
z=ax=<1RR(!6vN;r0+1ifq4kG`Lx<4i(8E*o|Pj
zD5%iwYAR6AK$U`Oa7?y@xWn@pQ8TYDoe2&NDXcD*A9@`1&`tfeQ)cgmdXkY3l|ts@
zUQDmh$nLD_?)?{CE!t#y2(_~TcI(MU;wG@6sgvuVNX2Q(UM4K&Q|*^RlI+`gsS#{lSq4r9tQSXY{Qf2IN)2
zNUU5ZR7^CsuZo5%)>cp9ZlIzYX3`sGF5uXL`mLHl4wby*5Y62g+~0c3ai(
zWlNP`>)dqUHskGNYrJry1V|_$mA(>3&EO`{TG9R780+h4`g!o?kgYl9;%R5HsFTvJ
zq8noE6pm#*ZuCg_HzvYGB66!~RuZU-4M#@$s6H5a-m2q>o61ZK<~Mb)<~%Agzb{*Q
zRwyGG?}Qa1!y#vS@|Zr%V(1(vnFnlM7^58uM>-wEnhoAcsWPf+wl(>PByg1#!y2Ca
zmnkCu{}+B6o-|Cqg6O>^G|z37v5rm4Yzg4T?qCxNQ+SGR9f#eQANfJ*Wft3!%{o*k(Z
zG}S>Z)+~48)6XXgVs3v4THO_DKJ0p*u($i*Xi`bQ4~TgQ{-#T|C)bBJ+2lZ%7TE2k
zTCBvqlTA$9cUi$cowJ+o;S1~MXeRIRZI|lxq*=`Q9kk^k1!T
zKJ}a8Fe3uV3%o5=*i_2@;_f|zntIrFUl0YPic+OS0Y#cpq?d?D6Cu)jlPmQ6IWx}(9A{WeU}de`_qy-v`dtj5
zgA7;@M|y;C`}UzKx{0E9AMhFPlnLPqVQ_DEMczdDRS)O_THnl5T!zV($MNyayDvi{
zl^c9;u_Zrz39&VwUVJ2>qZv8aydE9fA~cQCMhx`Hm4Q|@SKr0cIns+WkRM;G`3*mE
zbp!7mflI`#2kO}2*z4DT9RJblSJjrWGe4q_^Q2E|XYdIwf`u<$*B;MvlK-e5rgv8Q
z%$;x0;)&Z#8?dBYeNg6lk6KeI@HEc?<0+D@L%8a`hTUGC`rXzjANu2ucbO{x5-qky
z)lO2T|E+kN2uJoqk^Nx^aDH9I9u#pp%1dF9mw!XPQQeJTYD_|{o@B-xtjnX>7xi%H
z$Qj*_VSR}OIuYMGdgJ=!n7#&Xd0nfS8gW;^*AB>c!g)vVDYMwOK#vZo@a@{T~3YMP%pm?DnQZUmwnO;4mCGHh%A<
zyRFh*f720!fo>~m7I!DbNZAiQ>f`_j|Cn;W;B;0=Gnu#ME)kB7@@uA&OC`#+rRoo+
zB^u03kGD1LGzF522ZOhG-q6vuZ%5UL&TbXis?@YCP=1YvGoY3;!0!ufXNR_yfmaao
z?3NRNv-VQ8#eMu`$z*}=HGCKA%Tcq}xA*f6hpfhY^GNSy>Xm
zJo-N8x5lcGD$IGZsc}x=y2}CX@U|i&ybs2sfD})|aq^G2nJkg~vp(4(N8rA=J5F!{
z8Uor_Df=K&K7pIpl^iCMnD2DrlQuqfNCFl-Pcb?4-3%Ka2>b>=yZ6NvtF54%^bq-Y
z?3G%Ql8P^1foh?UOZ!^>D*8Of%_s9@&qSZ2m-9-WF?p3y-gcw%?|8=F9XXbuRf+oc
z;We9I)LiE$^7u$KBr@U-m|s)7aFOjg%}_hgx3x2I6|+pEW-z$(l{OZ*%o
z42)(n>(T~dY9`W1EaT_Tz?(<;W;IO4^N!(&vU<$@(hMc=-nptRFbP7i9!a!eKUGS0
zUp%xmOhK(~fgiFj%npP+dT!g}GUM%nC}qEstRDX-U7{5{d{UBP`zeaIDu&!JJ9Yx=
zmyjf=K$IoE3NO>mUZ`eEswZvIeikipJnS+t5k)i5k4Dfk04}r~VcFvXybTMqvJ$Zc
z!D6q=8IF^dn-3osy9LL5O#NmAC`ZAPSm|r#8w>$>sJQb}5iK9O=#uOooT1{^$L}&(
zWEij4^T#fvKJop270a^dl8ylO)R86j2F@Cv44eId$AS0IIa+jK9qsOjx}9VGeY1Ih
zKOkMcRQ^lEect;FJe0y$xOSh4AdeO6Mv@ddy>C>a7H2$|;0)&fSKzC5oUk6ov?*=#
z&{$oG`!|XfbodDK`C()0dMRWCnCCsMAOTN3^hUAW27rGPDgAA`0_E3OPeV=r@&m1?
z3WRF(hycHZE;S{XT^QZP$knL|Pinl&YjjaA!<#T(cwg2JmSgnzORJZACFV`m-tdAt
z?85)|0aEHy+|{d#U(VC;vG8yD=4Nnh$%~98}JU=6W#^#vg&URAWP!zZ6ItR5cFdS$u%O>?tyW#_Z*Hl
z`!?P%p$${T@?T%AZT_176SzcM=I=VQknmNw7XeW*96h4xid(#oX}Yj&(g=AoP_F
zXYJmKZsk8XCXAGR2x?t;yHRLEWN`FfJ4IFq#SP33S
z&%27s5Y2f;m>T~uc)5m&w7KO7)JaY)$6YsG-TeMkG4;8{X{ug+R-&K^+m;_xF_!VCdcy9Rbb3mjTa7AVx
z$+upb9IbYP>}D|W#roD@)03(VOew{ylvZ)u8#2LtkyFvgI@322OIRX`8<9)+IqXt#
z1x9cwJ}u)53T1AsYnXgVB6xoEEdB4{r@{f}6WOoKyaQj4GLs=UPyL33pq1opH^aoA
z{hq*)HT3&ZXFe%*cC#<+*5eilXDPY5vNfy}9#Xw+59w*<>}g$&QuTsX(S+gi15jw+
z(SLAyIXSNL{stGUZNZhAUQTxNN8JX4Ui%-l?)$Gw<$Fdc3g_HoUMW|;&n9ppnC1MX
zV`kH%tQkm7CH-We*rjzs(|7faEF=%VDNILvFUVv9*nMp`8V+QuOw!e{Wad+?r83ud
zMnxj8Cw~Us#KiF|=r|X$gPZg%B>V+V7JM28TgPX&d~3_aEBX7YwC$ogDy0BmQ+M0(VIPs{|4uLAN$$>
z@F4R$s+^`pzgOeP2!~CmHT_|rh))#~bETJS#g7C)Bv%!&w4m>KcCSX?F!D!%#mCRn
zEJZ}Rglyw^hdCrJf%414Q1tLjZ1`6<`-Y4yM|b{$So{ZeI4#}lF=7VXc;q-LOK}Hp
zh}b1S)Q$34Ysv+W1m9;X*BQMedbtUr6cd>%XYevBfDZb@Yp2U(AOw2jiFIS{eY(Oj
z4W7BH9vU?RI@Cy>OsO}(Ty9=pk2TnDl@PMTwqIo}MYl}`mD@z8Xnnf2Cs>*)Mm!Sa
zW$FZmj$TC{w@>FEeKqiPi3M)t=fH5|n|g3NRpsERr8qBms@e{eBgVZ%SE}|ZW!F=JYYGE
z^03abwUkp?Oe=bwWeW_^m95B-r8mCox&?9B$9%RgvyfC}9#Kk2(2)C(kAnpk`_X&W
zjU3{mNT3>WR0u`|;(rA6wyqC2v|oGJ(i%nQGlv(=B=o7>FrP%?;mziRE4#CpwCy+2
zmRO-?@b8AUPLvXyofpgw+u-E)NO(N-s~d#yHK)9P{>A$Xrl}mv05d^M70+#t+vD)z
z!VcE1Q9~SPZ{LaZatXaSFcIwLrM#A(#**;X{r#oqw>|5NE6UWTL?+`dM?%2ULPUH<
zYlwH&OW65eqL+y6dM~snQpeI3sq!sWz3j~!ya{K1^mc#df?kU;F|*}8+6TTzPRO0l
z_WQ0z+(SyPSDBg$tCq9~M$serT*T*~0An`Wb2`TQWvsyb+;)e+b8TI={l+L!igDOb
zvBJs?=dOqO9~V>BW!!m3Fxf=?QG-d8gd-3?a
z0+XUWX-7YjVa-=gIk`~tsF!3kJG=(z&wK*;0Q7vNxf0>)d5OF1!YAc&hUpEIO8kA$)cz&A_}XO3DwYSB8b1v9D0#B(vQ2-qIpPA
z+imFM)WzEIh{OyXyS%~!TDH5mmlHWTb+{xSzEl?L*$WvjyB!XfKm=<+9KxI9N+mTr
z)jr|93HEt(OpD!BNJtcwG;i+}M0hps4a*j1qebxFcbb@4D7EKd0Dn~H#Dq{PX-QY(
zK;Z-5JzptW0%Wvqj(`uZk0j#f4)=yxLT4?_2!rWT+4HJW@`IjGM*>oKR3~eiH`2gm
zaC(NG|H{_wg$uzk?f90L`8H9~S$TDYxB$4i+a^U=?oB_kVG|2&7jI)yFH!xgVYHNIx_XP69?
z_V2riy~E?E{K+^4oaL8JNJeZ8r%t({=LgHYRsWB#t%^DydYywdQ^yr&;w!}C<>CnM
zx~?)ts#_S!%i!$}$^_OW5->W!u+6+Jw#%z<6F~IB>g2I$riUnFyB?!s4cSQWJ?aB<
zP)jpOI{0N{8Ljkk+BL(DRaLWx;TO
zD>&D<3JF^70Y-c2*-<*>!Ioxu@X&8NfpM-L`Q7&Cuyc0in^vKlYzp+&=|*(OlZ(wj
z^!8a5h2EgN9zAVFGHVnz!C@dsohI7!i~vufaeTE-f#8st1@lV^4eiZ0hkMcPLwPKp
zLh#dpnexcbzOdZ``Rto0Xu3w-n8grPvM-O^`fOHbXXFgtDey2+RC__}m`QBjz`lBt
zWiaBMe4@P|9n;?m@-M6uD_bO``3p^OvHQFkL4{su|`CBAwW66mit-O
zb20xR#?ak;}MeqUmYn&)^|7
zbKIKS78WGOJc%QOZ3Nsc|2g)1L$CpxfpjV5Ty~H~2JyX=g^VD_Ku2n=-xV{l-9nmo
zhWCw*;&-M*wSI2w-3TIVhWY80(U^FP`RKEhVOr+etrUhbGzGN1RufF{#rJQ|dL3Y2
zFO$m|e8iWm$*Jd>$oGqMF5IxckxQ0}2bVzH2u+Q|S`A~mThBY1<*9I?j1>W|sA$Iq{
zZ27;#V*Z`2(cs%DqJ$*SRL#@;gX6FZG*-)%vn4;kRb}5^&%ELuSE)Nz7kV&LUHCli
z#oj@poI0D6y7ecL&aY-4PCCWmXbPit_>|<`%k;m9xP|I{HPx!Il~A3g^~5RmXR*nq
z_G9phd+q-+m`DhrsnwJ!>oDnPmi%5z`|AXFSSCF9__x|?Ck0E_G3I*ufJPY#bB0Hc
zEe|5VJffgW){Acr;%qm}k*WuJ9)W*!_YaD!%B!g5f5=$9`*y-8wpPCOcuhirE?ms!=XwVjqj%%_k;$ELGHu4eU7ZTo40GBul(eUPU
z^SE419n?F<8wk$h&AaGN71=LZ4~-Di0o~OC<+IJh)jmnFsdusRB>}R2S!f$s?k4;yKbcopookPQuYcymknALa_
zHlA0))1$X8RpagLm%u8RE1f88q&9Jz+L0L+Q~f8-Sv0Y?OMbBWo=V%_Tt9zGV^b=L
z=F}&oukKF(nm3@;`o$@2`m9_2CZk!h#@ZxmX=5Vn<+Xf2t*1abVoxVZ1l3^7iq9{O;Z+oO4B36YE-po;!{}#bnEaYuxpNd`kOX$
z8GDAoce;Xx7Ese=n|u#oD<`;~m^zn|l~;vGHiI8mob%UFn?^NyAx;Y#l(&i=JWo5x
zzZXD$J`7G!Vu~jLduvEw2u`sH_Vgd8N#>b^+e+vqBxmGJnGRe|LD*&`AeVa}*AgxQ-6Yp{u%$?Q@4XKF#(at70QsM3@EP9zNx_1LJ?ZVD~=7`#~W@s7(%9)&+<
z4c&vD6JjvHqk1Dp@FXyUW8BI^^SDs%;V{ltNlYhGUK$ph?nCcsK3O>xCTHllR0`m~8$?Tm
zw1MsLX4zT6dDw;Y`HN%PFGA$08l8F~n>J>qbaV`-y?fRpTLJ75Ry^s_RCoou$QB!9
zX%pDNi-m(ZZr*+sXSLMVFC`Bi_;FyM~*ZQIqbx1j?zy%u*-
z#&3p4!p`-|?#~4Ie|`woxS-Dp3K2~?GZa(MmP
zg{T?7Fey=~q{9!>Ocn`WoYGIHiqbsFS!Ml!yHS8^cc%dnXh=VJEZ>FOW@{Futn6Dc
zK3i`9>8yO0W7fwD?rPg1<2Rs)YaSQgR5};-!917G8d?YRKUNXPh!0bWx_`{+)j544
zd~J%Gt6$b}M*l?-6s(R_l(T0n8(IGzi7KSg^vETq<
zj=QeiB?yOs^Yi?lil>zy?<{!Px}$d$8x81Qo0XE^z2|Vq&8KKF%%SZzsFO+nx6qL#
zTt3D6OxT<{7Rw3)cAKJx=25`|t&<592>|Vy8WlWRmGB9JGX%Oni!|^ZBCp6DWNeCT
zn?bRGU}Hf(I&u7a7sQcSEaQO~_K3Zdhcf;nM{=uYMTW{fSS>^z6j5?V#GA1+FzOdD
zjRrPYL_6g@jZSrNL#4O2v(%D$c8u<|I9Q{;QU#(~p5$v|P~W#c|2lKOWMDrOrxm{u
z5lZ8PeG~yZy_^Jdn>oZMg7lAdonQJ{O=nC?K*7g_ykq@>
zx;WtQJP7}i7kdd6HLwqSD_=bM1Kk#_%wb+(O!zTt%Spw{cKXG%gfCP#k?;WiH6;3w
zfe4@gb16D}TD0Qw1WE|Vd*0f;19!#SRAlF#mVZ+MlLy-oXtUYS-?YKJO%;}xBSGIl
z>?W6urfQ*)%k@np8*t`Tzt5jWvQ;jC%xw6^)2*Bv#d#$;+Z5}-r?(bX<9vHYhIo5m#dF~H7vWp
z%%qXFFLoLUHp&OQL`MJ1ON4I8+`^TiRK@un_Tt~}tF&mPwd0T)ZFKyOiSfuNqztW4
zJXnoM?RbdQrO`&u&>wWsTD}<*Q8(yxqUuBL(4Z;Uur@-NcIY|wt0?U9IRubQz@RJK
zmo&%9{aC7th-XtYnq!}kIXB?**7VB3tXi($B2LdLZLQWE#0KrYexF9V+_+hRj%EaB
zwb~tmMT3H?{@T%KJ)UYV@S3Y0Boj7ifIl!8ru~%k@cmxVog#Vh0=kYD^T1tj$xRSw
zd)PvDiPz<<=-3dIRHl&A{;=}9E@zQPqO7s32YzL^H3>f5$$b{aCG_ZB78fN^bvAxhLyUo4uBSFBqBU;+%
z@R#^z_k8leq6Ti28K4CZU<)GVwWj*OMy5fG4`Tk?B9F;MD>#&S_HDhndM7NtNc?MyU|)J>tfX3F=T0}4+*l?KEsnlp>#8|25K$`T$I`gM9?CD
zaW`?DGI+L+cZaj6!NiU3AsfqOT`~Q4yU|dKA@Wj!ZP}lh$$mPT!eZEf+&b+w+yyB=
z%;{O4fq@JSMMPAZlg`+O6H8dqE|Re;%wc@
z%R*cYw5#oc_!fsJVO9%lQiy_0R6!P;O!naOJkPgH5ac+gq%@dsao|@!NmQe^MuMs*
z9vfo_74y@W6)*eNwTZ$Oqxu~PXp>3^7)4TamzV0h%s{&{Q#Dh*;(`0pITP)&ST6(w
znmAg;$iU#uTkK>{6g{;)garwnDw8IX>p1bYECIbv{>TI7NKh}sOWfX9^R*9qrUH*g
z?0o)XjNku|VB@oJRyRa#DlbUVw&4)8z%Vv;A!83Aq9b5nLtdp;Jq9=xjXwcN=n%}o
zVbi#7x+Sp0T2Hs|!B7?k+qOXEq
zMNMJHMNXG~V@x09$v^DVl}5T=_~lul7~FvoHBq93u=H?0Yyj&D5W-6v(>k}E$~)dD
zY{$vxYjrw26>%=Yj3B`2?o9e$(Djm?;JzTZ#Z8vg1>i!%Hd9XxXo&?GSrp`QXX^{21_GzQ*9?Y;PBX
zkQth^2v}cqm1zuMV93_b-)TH5=Ay>Z0O%?%Tq&wKW?}Sg(TV3gM>%zwumCecQ4Ek8m_Xt~k*PFU7
z%Xv@Vz67-j5r*|H;_Chu^P5<#7bRQ6F
zeV_GYx);NHHDyuu>knMz-FGrS3p6<>1(f4a^0{Iza7f*3wN#-O1KHKtRJ4As!^#j3
zGyH>l2~~T4y+G-BxFY=RnNv+g!tP0&tZ=EtW+%HH(>t>i8m8tRI`r@w=Zt+ZFKyd7
zI70zlbZ1i$pYo~SR~jTkL2-AP^eN)^`l}~^NT`A
zNzD5T6$ByTL((Md%t^>#<|Ag()=}q(;6~%sTr@ES-m$-H7DLVF7;W-{H9YS&`5*TA>ZF$oKZ`$ZI
zo!b^xo~C6qbDj~Lt4x_qeW9^oj}N+XPjhEgYyl!+rv)D+&1Tg9gOkKL3;cW67J=rc2?auX9UFkR
zo%1mqSj?&)1s#R~$P~50IYkEsLnb6wjbhD@a?8+fwO?iLq`_0%=+^vev^RzElh72^
zmglBGgg_LH(Y1`ZZ1aaL6b6I9`Kd!k$%Itw@+K#Bd|WRxwX;n@3ftI?P5p=ONW
z|I$(Z7Z~Ti>nKHAFJFN74*xFA`mDSP{sLP52dBL2b^$-}@@0)!AZc?G#mqv1J4Z3%
z6|2=9PKLUzUkPKF2&je6@Ql!M^Kg&=w>K$NrDfu$R(;~sbQj6PUWdC5MAVc^
zUyrVp(@Bs^0fZc+4CV6_y{Dh>jtAd<;(PU~Tw~^y5c|mp&_bzOXU9MJN}vmLd~t#x
z9=tr|RaAlMqc-B+E#1c9%%0~iq&vM5xewmLxP#-GYE{l0ZS$3UNBR!-(g;fjs|H2
ziebWb<`D7TxxzFmPy~f$2i%_1`GGF7O{(G&AHpJo{us}-N&-D@9I953qjn?J@I0LFN3?AIXkeV4||F$9(bcR
z(HNUljfd`*w(N#Io7EtKVaNXD^E-YQ&|!fcKe{6>p8Z$UtEeI9@Dv;cCd(Gqi2jo9
zUdC!!FR)-O%KAMsTSE%Dv39)*+v1E;YpjwWRF&@^LyJ7-@MnTt$bo*9<|&ONIO_69
zHi0&=RAr{0+(|YnFLZs$2#!#sEsGe+b%^`Y{Mtq$P-PE}CbDtCGf2H=AZ}kaz+51{!gc|I0E;X{Y)aL?~z3;=<
zOjIcY7dM8&KIgE~=k&H~U~6*w%nlK&A8V?3@t%;SlgcCcDA|Gq
zgh9hC!u}J!q6-F`qm;1MFESS?z@l21+=eJy(J!t470M{|(nAo4XaO2)W}7d;cU{JF
z%nheQ(}3(lh$$AZ-Sd@Yu4X(rs^q*&{LQI)dD*26<8-Bk=dv<_q=~)vG(jbrZU`8L
zf%=+vb;qfaTyWkeh_FcDE8MEEI}EIiZK8ST6Xk2k+wj4euh}Jk{Z08R^jJhY<-Ktk
zL2KW%c(iGBUmcLF)n&iP+ctDS!)EUYh3y(ie`cZwS*3f2M3jx+{R%Z2YXVOF2qY~B
z-%OKp-Ww3nVmJ%=wHib?)JBf&q~D)RWQWF&td^}@eyaC{C|c>4U&#TSjq#&0o5|Od
zKTTMy&DwhI1`zfj@>!8oMOGUkcU{O6>W2x+90rKCBPRET?vIOa$p5r9X*1L&OVGAU
z;#iT|Z@(Szlo@Dj#GgCM$!vKWN9VX)2wJf7hHXbAZms>`vRKe5?~%^9l<+<;e8><}
z)-8q-ccrS$7PYmfukt{(3&1?=5;^|C;oYI%UK03xlST6oHh7^-?&vc3y0-g?qt+I*
z?7gfSUFQk=ZfFb+&+;;ZThZ9L8Clqhxj|X)HWHOPkjuy0aehWcEUVJ9h50PMLE65I
zx}{$|moR(uf*_cY+`%@iNHpOSZ{QS+<4W$yvfBZa$e^D&4VbJ?O&P0OlXJ+D0h#gn
zQKk-lO;C;sCo=WG-hK5{2zOIGzFadvvU(_~p+i~AuQ4mR+m2eB*um;CZSTT1m1C);
z09Y2#Y*Xs2DF-s4A4d#(c*e-FpKY6730b)f_xV*Bgl@aP<6E@VI&;%S&?(6^zOMS9I;`m!Np(6D
zNs?h~sgK)(-W~mdmrwge#(2H5#TlOWK^;(HS=}gNiL8$aty(M62Y=6C@;$TEE5-&E
z?T4%E+?!5$s}F}~1nmH{l}_4mpb?gE4%qef{?A=+U&E`Yr7FX$8*RTKQLP}srx?8JXw#&SzA5vYc&8b7c((8U;g~%30YzDclC*;zszNP
zujHdKj+SV+zIh);NG<&5yu3)1>o-PQY#;Qbdw<5Lua(Y|05IB+xye@p;l&P`M=e=g
z%zwosEV}g+GMIIwzr{*%=+N4C^W}|9T>B4?{(CASaj$`H_%3jlrw;#v<7k9w1t5aN
zkmcxo@?a2|xXJkfY(})J`_}Oy=`g$e5OBP750|h0C$ez8br~o)|3ye#gtl8YL2_r&
z_mAX4%UMN8xN8Npj`a9Yi~r_87C3MH
z4w>1rh)KJre^C|(UXI0*HXkAO;!+bPYEPv8O88xTzB;C0p<}JlUTvT3AGQ$7W5!+uHWvr`7OaMhX=Qf~dDJqWdf}
zv|lXe@Wy5nqpH!w#b6?JPH+V1h{NQ>#5Xp@M%dAWb
zHOBz{hUr-W!0{^((R1oxXT$upcK_`>CnIF*nClH1P**94h4ZxS5S-;?c}R0nbG3wD
zsXCJRTz~XAKAcTwfC_GGfgh)