Pedestrian Attribute Recognition Method and Recognition System Based on Sequence Context Learning

Abstract

The invention discloses a pedestrian attribute recognition method and recognition system based on sequence context relationship learning, wherein the pedestrian attribute recognition method includes: establishing and training a pedestrian recognition system; encoding an image to be recognized into an image sequence in the vertical direction, and initializing the attribute sequence is a random value; use the trained pedestrian recognition system to calculate the image context sequence and attribute context sequence; calculate the image context sequence to the attention of each element in the attribute context sequence; calculate each attribute belonging to each category of the attribute The probability of , select the category with the highest probability value as the category of the attribute. This method makes full use of the contextual relationship between image sequences, the contextual relationship between attributes, and the contextual relationship between images and attributes, and improves the accuracy of pedestrian attribute recognition.

Description

technical field [0001] The invention belongs to the technical field of pedestrian attribute recognition, and in particular relates to a pedestrian recognition method and system combining image and attribute intra-class and inter-class relationships. Background technique [0002] The task of pedestrian attribute recognition is to predict the attribute labels of pedestrians in the image, including age, gender, color of clothes, etc. These attributes contain rich semantic information that can describe the appearance of pedestrians, can bring beneficial information to pedestrian recognition tasks, and have high application value, so they have received extensive attention. The main difficulty is that changes in pedestrian angles and photo illumination, as well as long distances will affect the accuracy of recognition. [0003] In order to improve the accuracy of pedestrian attribute recognition, most of the current methods enter a whole image into the classification network, and...

AI Technical Summary

Problems solved by technology

Therefore, some methods use manual segmentation of images or manual grouping of attributes to obtain the contextual relationship between elements between images or attributes, but these methods need to use prior knowledge to divide the image into a fixed number of blocks or attributes. Divided into fixed groups, such a fixed grouping makes the learning of the contextual relationship between attributes limited and somewhat unreasonable, and the universality is low. When new attributes are added, they need to be regrouped

Images