Target object detection method and device and server

Abstract

The invention provides a target object detection method and device and a server. Based on the method, before specific implementation, a preset processing model at least comprising an improved feature extractor is constructed and trained, the improved feature extractor at least comprises six convolutional layers, and the sixth convolutional layer in the six convolutional layers is used for extracting deep features of a target object; in specific implementation, a preset processing model can be called to process a target image in a target scene, and a corresponding target processing result is obtained; and according to the target processing result, it is determined whether the target image contains the target object and the object type and position information of the target object. By calling the preset processing model to process the target image, the deep features of the target object with a small area range in the target image can be effectively extracted, and then the object category and position information of the target object in the target image can be accurately recognized by combining the deep features with the shallow features.

Description

technical field [0001] This specification belongs to the technical field of artificial intelligence, and in particular relates to a detection method, device and server of a target object. Background technique [0002] In some data processing scenarios (for example, bank-related business processing scenarios), sometimes it is necessary to analyze and identify the image data collected during the data processing process, and determine whether the image contains relevant target objects by identifying Whether the data processing complies with the relevant requirements. For example, when identifying and determining whether to handle a specific business, whether the user has provided complete documents, documents and other materials. [0003] Since the area occupied by the real target object to be identified and determined contained in the collected image data is often small, it is difficult to identify and determine the collected image data based on existing methods. Errors are ...

AI Technical Summary

Problems solved by technology

[0003] Since the area occupied by the real target object to be identified and determined contained in the collected image data is often small, it is difficult to identify and determine the collected image data based on existing methods. Prone to error, low recognition accuracy

[0004] For the above problems, no effective solution has been proposed

Images