Pedestrian re-identification method based on TOIM loss function

Abstract

The invention discloses a pedestrian re-identification method based on a TOIM loss function, and the method comprises the following steps: S1, preparing a data set, and enabling pedestrians with different identities in the data set to have different IDs; S2, establishing a network model based on a ResNet-50 model by utilizing a pytorch framework; S3, inputting the pictures of the training set in the data set into the network model for training; and S4, inputting the pictures of the test set in the data set into the trained network model, and identifying the pedestrian identity. According to the pedestrian re-identification method, the advantages of an online instance matching (OIM) loss function and a triple (Triplet) loss function are combined, meanwhile, attention to difficult samples isemphasized, and the batch processing construction process of the triple loss function is simplified, so that the convergence speed is greatly increased, and the pedestrian re-identification accuracyis effectively improved.

Description

technical field [0001] The invention belongs to the technical field of pedestrian re-identification, in particular to a pedestrian re-identification method based on TOIM loss function. Background technique [0002] Pedestrian re-identification, also known as pedestrian re-identification, is a technology that uses computer vision technology to determine whether a specific pedestrian exists in an image or video sequence. Widely regarded as a subproblem of image retrieval. Given a monitored pedestrian image, retrieve the pedestrian image across devices. It is designed to make up for the visual limitations of the current fixed camera, and can be combined with pedestrian detection / pedestrian tracking technology, and can be widely used in intelligent video surveillance, intelligent security and other fields. [0003] Due to the differences between different camera equipment, pedestrians have both rigid and flexible characteristics, and their appearance is easily affected by clot...

AI Technical Summary

Problems solved by technology

But in the process of practical application, it is found that there are still some defects in this literature. For example, for the cross-entropy loss function, as the number of pedestrian types (people with different identities) increases, it will become extremely difficult to train such a large cross-entropy classifier. Slow, while the cross-entropy loss function treats all samples equally, or even worse, the network will not be able to converge; for the triple loss function, no difficult samples are selected for training, and when the input data continues to increase, the sampling complexity is also promote

Images