[Doc] Add model zoo and update doc index (open-mmlab#1618)

* add task-level README files

* update README.md

* update compiling commands

* update doc dependency

* fix bugs

* update cn readme

* update cn readme

* update sphinx version

* fix bug

* modify doc structure

* fix bug

* add cn doc skeleton

* update cn docs

configs/animal_2d_keypoint/README.md

@@ -0,0 +1,18 @@
+# 2D Animal Keypoint Detection
+
+2D animal keypoint detection (animal pose estimation) aims to detect the key-point of different species, including rats,
+dogs, macaques, and cheetah. It provides detailed behavioral analysis for neuroscience, medical and ecology applications.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_animal_keypoint.md) to prepare data.
+
+## Demo
+
+Please follow [DEMO](/demo/docs/2d_animal_demo.md) to generate fancy demos.
+
+<img src="https://user-images.githubusercontent.com/11788150/114201893-4446ec00-9989-11eb-808b-5718c47c7b23.gif" height="140px" alt><br>
+
+<img src="https://user-images.githubusercontent.com/11788150/114205282-b5d46980-998c-11eb-9d6b-85ba47f81252.gif" height="140px" alt><br>
+
+<img src="https://user-images.githubusercontent.com/11788150/114023530-944c8280-98a5-11eb-86b0-5f6d3e232af0.gif" height="140px" alt><br>

configs/animal_2d_keypoint/topdown_heatmap/README.md

@@ -0,0 +1,7 @@
+# Top-down heatmap-based pose estimation
+
+Top-down methods divide the task into two stages: object detection and pose estimation.
+
+They perform object detection first, followed by single-object pose estimation given object bounding boxes.
+Instead of estimating keypoint coordinates directly, the pose estimator will produce heatmaps which represent the
+likelihood of being a keypoint.

configs/body_2d_keypoint/README.md

@@ -0,0 +1,19 @@
+# Human Body 2D Pose Estimation
+
+Multi-person human pose estimation is defined as the task of detecting the poses (or keypoints) of all people from an input image.
+
+Existing approaches can be categorized into top-down and bottom-up approaches.
+
+Top-down methods (e.g. DeepPose) divide the task into two stages: human detection and pose estimation. They perform human detection first, followed by single-person pose estimation given human bounding boxes.
+
+Bottom-up approaches (e.g. Associative Embedding) first detect all the keypoints and then group/associate them into person instances.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_body_keypoint.md) to prepare data.
+
+## Demo
+
+Please follow [Demo](/demo/docs/2d_human_pose_demo.md#2d-human-pose-demo) to run demos.
+
+<img src="/demo/resources/demo_coco.gif" width="600px" alt>

configs/body_2d_keypoint/topdown_heatmap/README.md

@@ -0,0 +1,7 @@
+# Top-down heatmap-based pose estimation
+
+Top-down methods divide the task into two stages: object detection and pose estimation.
+
+They perform object detection first, followed by single-object pose estimation given object bounding boxes.
+Instead of estimating keypoint coordinates directly, the pose estimator will produce heatmaps which represent the
+likelihood of being a keypoint.

configs/body_2d_keypoint/topdown_regression/README.md

@@ -0,0 +1,5 @@
+# Top-down regression-based pose estimation
+
+Top-down methods divide the task into two stages: object detection and pose estimation.
+
+They perform object detection first, followed by single-object pose estimation given object bounding boxes. With features extracted from the bounding box area, the model learns to directly regress the keypoint coordinates.

configs/body_3d_keypoint/README.md

@@ -0,0 +1,13 @@
+# Human Body 3D Pose Estimation
+
+3D human body pose estimation aims at predicting the X, Y, Z coordinates of human body joints. Based on the camera number to capture the images or videos, existing works can be further divided into multi-view methods and single-view (monocular) methods.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/dataset_zoo/3d_body_keypoint.md) to prepare data.
+
+## Demo
+
+Please follow [Demo](/demo/docs/3d_human_pose_demo.md) to run demos.
+
+<img src="https://user-images.githubusercontent.com/15977946/118820606-02df2000-b8e9-11eb-9984-b9228101e780.gif" width="600px" alt><br>

configs/face_2d_keypoint/README.md

@@ -0,0 +1,16 @@
+# 2D Face Landmark Detection
+
+2D face landmark detection (also referred to as face alignment) is defined as the task of detecting the face keypoints from an input image.
+
+Normally, the input images are cropped face images, where the face locates at the center;
+or the rough location (or the bounding box) of the hand is provided.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_face_keypoint.md) to prepare data.
+
+## Demo
+
+Please follow [Demo](/demo/docs/2d_face_demo.md) to run demos.
+
+<img src="https://user-images.githubusercontent.com/11788150/109144943-ccd44900-779c-11eb-9e9d-8682e7629654.gif" width="600px" alt><br>

configs/face_2d_keypoint/topdown_heatmap/README.md

@@ -0,0 +1,7 @@
+# Top-down heatmap-based pose estimation
+
+Top-down methods divide the task into two stages: object detection and pose estimation.
+
+They perform object detection first, followed by single-object pose estimation given object bounding boxes.
+Instead of estimating keypoint coordinates directly, the pose estimator will produce heatmaps which represent the
+likelihood of being a keypoint.

configs/fashion_2d_keypoint/README.md

@@ -0,0 +1,7 @@
+# 2D Fashion Landmark Detection
+
+2D fashion landmark detection (also referred to as fashion alignment) aims to detect the key-point located at the functional region of clothes, for example the neckline and the cuff.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_fashion_landmark.md) to prepare data.

configs/hand_2d_keypoint/README.md

@@ -0,0 +1,16 @@
+# 2D Hand Pose Estimation
+
+2D hand pose estimation is defined as the task of detecting the poses (or keypoints) of the hand from an input image.
+
+Normally, the input images are cropped hand images, where the hand locates at the center;
+or the rough location (or the bounding box) of the hand is provided.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/tasks/2d_hand_keypoint.md) to prepare data.
+
+## Demo
+
+Please follow [Demo](/demo/docs/2d_hand_demo.md) to run demos.
+
+<img src="https://user-images.githubusercontent.com/11788150/109098558-8c54db00-775c-11eb-8966-85df96b23dc5.gif" width="600px" alt><br>

configs/hand_2d_keypoint/topdown_heatmap/README.md

@@ -0,0 +1,7 @@
+# Top-down heatmap-based pose estimation
+
+Top-down methods divide the task into two stages: object detection and pose estimation.
+
+They perform object detection first, followed by single-object pose estimation given object bounding boxes.
+Instead of estimating keypoint coordinates directly, the pose estimator will produce heatmaps which represent the
+likelihood of being a keypoint.

configs/hand_3d_keypoint/README.md

@@ -0,0 +1,7 @@
+# 3D Hand Pose Estimation
+
+3D hand pose estimation is defined as the task of detecting the poses (or keypoints) of the hand from an input image.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/dataset_zoo/3d_hand_keypoint.md) to prepare data.

configs/hand_gesture/README.md

@@ -0,0 +1,13 @@
+# Gesture Recognition
+
+Gesture recognition aims to recognize the hand gestures in the video, such as thumbs up.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_hand_gesture.md) to prepare data.
+
+## Demo
+
+Please follow [Demo](/demo/docs/gesture_recognition_demo.md) to run the demo.
+
+<img src="https://user-images.githubusercontent.com/15977946/172213082-afb9d71a-f2df-4509-932c-e47dc61ec7d7.gif" width="600px" alt>

configs/wholebody_2d_keypoint/README.md

@@ -0,0 +1,19 @@
+# 2D Human Whole-Body Pose Estimation
+
+2D human whole-body pose estimation aims to localize dense landmarks on the entire human body including face, hands, body, and feet.
+
+Existing approaches can be categorized into top-down and bottom-up approaches.
+
+Top-down methods divide the task into two stages: human detection and whole-body pose estimation. They perform human detection first, followed by single-person whole-body pose estimation given human bounding boxes.
+
+Bottom-up approaches (e.g. AE) first detect all the whole-body keypoints and then group/associate them into person instances.
+
+## Data preparation
+
+Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_wholebody_keypoint.md) to prepare data.
+
+## Demo
+
+Please follow [Demo](/demo/docs/2d_wholebody_pose_demo.md) to run demos.
+
+<img src="https://user-images.githubusercontent.com/9464825/95552839-00a61080-0a40-11eb-818c-b8dad7307217.gif" width="600px" alt><br>

configs/wholebody_2d_keypoint/topdown_heatmap/README.md

@@ -0,0 +1,7 @@
+# Top-down heatmap-based pose estimation
+
+Top-down methods divide the task into two stages: object detection and pose estimation.
+
+They perform object detection first, followed by single-object pose estimation given object bounding boxes.
+Instead of estimating keypoint coordinates directly, the pose estimator will produce heatmaps which represent the
+likelihood of being a keypoint.

demo/docs/2d_animal_demo.md

@@ -15,7 +15,7 @@ python demo/topdown_demo_with_mmdet.py \
     [--device ${GPU_ID or CPU}]
 ```
 
-The pre-trained animal pose estimation model can be found from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/tasks/animal.html).
+The pre-trained animal pose estimation model can be found from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/animal.html).
 Take [animalpose model](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_animalpose_256x256-1aa7f075_20210426.pth) as an example:
 
 ```shell

demo/docs/2d_face_demo.md

@@ -16,7 +16,7 @@ python demo/topdown_face_demo.py \
     [--kpt-thr ${KPT_SCORE_THR}]
 ```
 
-The pre-trained face keypoint estimation model can be found from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/tasks/face.html).
+The pre-trained face keypoint estimation model can be found from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/face.html).
 Take [aflw model](https://download.openmmlab.com/mmpose/face/hrnetv2/hrnetv2_w18_aflw_256x256-f2bbc62b_20210125.pth) as an example:
 
 ```shell

demo/docs/2d_hand_demo.md

@@ -19,7 +19,7 @@ python demo/topdown_demo_with_mmdet.py \
 
 ```
 
-The pre-trained hand pose estimation model can be downloaded from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/tasks/hand.html).
+The pre-trained hand pose estimation model can be downloaded from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/hand.html).
 Take [onehand10k model](https://download.openmmlab.com/mmpose/hand/hrnetv2/hrnetv2_w18_onehand10k_256x256-30bc9c6b_20210330.pth) as an example:
 
 ```shell

demo/docs/2d_human_pose_demo.md

@@ -16,7 +16,7 @@ python demo/image_demo.py \
     [--draw_heatmap]
 ```
 
-The pre-trained hand pose estimation model can be downloaded from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/tasks/body.html).
+The pre-trained hand pose estimation model can be downloaded from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/body.html).
 Take [coco model](https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w48_coco_256x192-b9e0b3ab_20200708.pth) as an example:
 
 ```shell

demo/docs/2d_wholebody_pose_demo.md

@@ -16,7 +16,7 @@ python demo/image_demo.py \
     [--draw_heatmap]
 ```
 
-The pre-trained hand pose estimation model can be downloaded from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/tasks/wholebody.html).
+The pre-trained hand pose estimation model can be downloaded from [model zoo](https://mmpose.readthedocs.io/en/1.x/model_zoo/wholebody.html).
 Take [coco-wholebody_vipnas_res50_dark](https://download.openmmlab.com/mmpose/top_down/vipnas/vipnas_res50_wholebody_256x192_dark-67c0ce35_20211112.pth) model as an example:
 
 ```shell