Create SegmentationResult doc and evaluation functions (#119)

* Update README.md

* Update README.md

* Update README.md

* Create README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Add evaluation calculate time and fix some bugs

* Update classification __init__

* Move to ppseg

* Add segmentation doc

* Add PaddleClas infer.py

* Update PaddleClas infer.py

* Delete .infer.py.swp

* Add PaddleClas infer.cc

* Update README.md

* Update README.md

* Update README.md

* Update infer.py

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Add PaddleSeg doc and infer.cc demo

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Update README.md

* Create segmentation_result.md

* Update README.md

* Update segmentation_result.md

* Update segmentation_result.md

* Update segmentation_result.md

* Update classification and detection evaluation function

Co-authored-by: Jason <jiangjiajun@baidu.com>

csrc/fastdeploy/vision/segmentation/ppseg/ppseg_pybind.cc

@@ -24,7 +24,6 @@ void BindPPSeg(pybind11::module& m) {
               pybind11::array& data) {
              auto mat = PyArrayToCvMat(data);
              vision::SegmentationResult* res = new vision::SegmentationResult();
-             // self.Predict(&mat, &res);
              self.Predict(&mat, res);
              return res;
            })

docs/api/vision_results/README.md

@@ -4,7 +4,8 @@ FastDeploy根据视觉模型的任务类型，定义了不同的结构体(`csrcs
 
 | 结构体 | 文档 | 说明 | 相应模型 |
 | :----- | :--- | :---- | :------- |
-| ClassificationResult | [C++/Python文档](./classification_result.md) | 图像分类返回结果 | ResNet50、MobileNetV3等 |
+| ClassifyResult | [C++/Python文档](./classification_result.md) | 图像分类返回结果 | ResNet50、MobileNetV3等 |
+| SegmentationResult | [C++/Python文档](./segmentation_result.md) | 图像分割返回结果 | PP-HumanSeg、PP-LiteSeg等 |
 | DetectionResult | [C++/Python文档](./detection_result.md) | 目标检测返回结果 | PPYOLOE、YOLOv7系列模型等 |
 | FaceDetectionResult | [C++/Python文档](./face_detection_result.md) | 目标检测返回结果 | SCRFD、RetinaFace系列模型等 |
 | FaceRecognitionResult | [C++/Python文档](./face_recognition_result.md) | 目标检测返回结果 | ArcFace、CosFace系列模型等 |

docs/api/vision_results/segmentation_result.md

@@ -0,0 +1,32 @@
+# SegmentationResult 目标检测结果
+
+SegmentationResult代码定义在`csrcs/fastdeploy/vision/common/result.h`中，用于表明图像中每个像素预测出来的分割类别和分割类别的概率值。
+
+## C++ 定义
+
+`fastdeploy::vision::DetectionResult`
+
+```
+struct DetectionResult {
+  std::vector<uint8_t> label_map;
+  std::vector<float> score_map;
+  std::vector<int64_t> shape;
+  bool contain_score_map = false;
+  void Clear();
+  std::string Str();
+};
+```
+
+- **label_map**: 成员变量，表示单张图片每个像素点的分割类别，`label_map.size()`表示图片像素点的个数
+- **score_map**: 成员变量，与label_map一一对应的所预测的分割类别概率值(当导出模型时指定`without_argmax`)或者经过softmax归一化化后的概率值(当导出模型时指定`without_argmax`以及`with_softmax`或者导出模型时指定`without_argmax`同时模型初始化的时候设置模型[类成员属性](../../../examples/vision/segmentation/paddleseg/cpp/)`with_softmax=True`)
+- **shape**: 成员变量，表示输出图片的shape，为H\*W
+- **Clear()**: 成员函数，用于清除结构体中存储的结果
+- **Str()**: 成员函数，将结构体中的信息以字符串形式输出（用于Debug）
+
+## Python 定义
+
+`fastdeploy.vision.SegmentationResult`
+
+- **label_map**(list of int): 成员变量，表示单张图片每个像素点的分割类别
+- **score_map**(list of float): 成员变量，与label_map一一对应的所预测的分割类别概率值(当导出模型时指定`without_argmax`)或者经过softmax归一化化后的概率值(当导出模型时指定`without_argmax`以及`with_softmax`或者导出模型时指定`without_argmax`同时模型初始化的时候设置模型[类成员属性](../../../examples/vision/segmentation/paddleseg/python/)`with_softmax=true`)
+- **shape**(list of int): 成员变量，表示输出图片的shape，为H\*W

examples/vision/classification/paddleclas/cpp/README.md

@@ -16,7 +16,7 @@ tar xvf fastdeploy-linux-x64-gpu-0.2.0.tgz
 cd fastdeploy-linux-x64-gpu-0.2.0/examples/vision/classification/paddleclas/cpp
 mkdir build
 cd build
-cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/../../../../../../fastdeploy-linux-x64-gpu-0.2.0 
+cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/../../../../../../../fastdeploy-linux-x64-gpu-0.2.0 
 make -j
 
 # 下载ResNet50_vd模型文件和测试图片

fastdeploy/vision/evaluation/classify.py

@@ -71,7 +71,12 @@ def eval_classify(model, image_file_path, label_file_path, topk=5):
     topk_acc_score = topk_accuracy(np.array(result_list), np.array(label_list))
     if topk == 1:
         scores.update({'topk1': topk_acc_score})
+        scores.update({
+            'topk1_average_inference_time(s)': average_inference_time
+        })
     elif topk == 5:
         scores.update({'topk5': topk_acc_score})
-    scores.update({'average_inference_time': average_inference_time})
+        scores.update({
+            'topk5_average_inference_time(s)': average_inference_time
+        })
     return scores

fastdeploy/vision/evaluation/detection.py

@@ -28,6 +28,7 @@ def eval_detection(model,
     from .utils import COCOMetric
     import cv2
     from tqdm import trange
+    import time
 
     if conf_threshold is not None or nms_iou_threshold is not None:
         assert conf_threshold is not None and nms_iou_threshold is not None, "The conf_threshold and nms_iou_threshold should be setted at the same time"
@@ -80,6 +81,6 @@ def eval_detection(model,
     eval_metric.accumulate()
     eval_details = eval_metric.details
     scores.update(eval_metric.get())
-    scores.update({'average_inference_time': average_inference_time})
+    scores.update({'average_inference_time(s)': average_inference_time})
     eval_metric.reset()
     return scores