Image object detection

EasyVision of Machine Learning Platform for AI (PAI) allows you to perform model training and prediction in image object detection. You can also use EasyVision to perform distributed training and prediction on multiple servers. This topic describes how to use a trained model of EasyVision to perform offline object detection.

Data formats

For more information, see Input data formats.

Based on the generated file list, you can start prediction tasks of image object detection by running the PAI command. You can call the PAI command by using the SQL Script component, the MaxCompute client, or an ODPS SQL node of DataWorks. For more information, see MaxCompute client (odpscmd) or Develop a MaxCompute SQL task.

pai -name ev_predict_ext
             -Dmodel_path='OSS path of your model'
             -Dmodel_type='detector'
             -Dinput_oss_file='oss://path/to/your/filelist.txt'
             -Doutput_oss_file='oss://path/to/your/result.txt'
             -Dimage_type='url'
             -Dnum_worker=2
             -DcpuRequired=800
             -DgpuRequired=100
             -Dbuckets='Your OSS folder'
             -Darn='Alibaba Cloud Resource Name (ARN) of your role'
             -DossHost='Your OSS domain'

For more information, see Parameters.

Output

Prediction results are written to an output file. Each entry in the file consists of an OSS path and a JSON sting. The OSS path indicates the path of the original image, whereas the JSON string indicates the prediction result. The entries are in the following format:

oss://path/to/your/image1.jpg, JSON string
oss://path/to/your/image1.jpg, JSON string
oss://path/to/your/image1.jpg, JSON string

The JSON string is in the following format:

{
  "detection_boxes": [[243.5308074951172, 197.69570922851562, 385.59625244140625, 247.7247772216797], [292.1929931640625, 114.28043365478516, 571.2748413085938, 165.09771728515625]], 
  "detection_scores": [0.9942291975021362, 0.9940272569656372],
  "detection_classes": [1, 1],
  "detection_classe_names": ["text", "text"]
 }

The following table describes the parameters in the JSON string.

Parameter	Description	Shape	Type
detection_boxes	The detected area with coordinates in the following order: [top, left, bottom, right].	[num_detections, 4]	FLOAT
detection_scores	The detection probability of the object.	num_detections	FLOAT
detection_classes	The ID of the class to which the object belongs.	num_detections	INT
detection_class_names	The name of the class to which the object belongs.	num_detections	STRING