Pedestrian detection method based on end-to-end convolutional neural network

Abstract

The invention discloses a pedestrian detection method based on an end-to-end convolutional neural network in order to solve the problem that the existing pedestrian detection algorithm has the disadvantages of low detection precision, complex algorithm and difficult multi-module fusion. A novel end-to-end convolutional neural network is adopted, a training sample set with marks is constructed, and end-to-end training is performed to get a convolutional neural network model capable of predicting a pedestrian candidate box and the confidence of the corresponding box. During test, a test picture is input into a trained model, and a corresponding pedestrian detection box and the confidence thereof can be obtained. Finally, non-maximum suppression and threshold screening are performed to get an optimal pedestrian area. The invention has two advantages compared with previous inventions. First, through end-to-end training and testing, the whole model is very easy to train and test. Second, pedestrian scale and proportion problems are solved by constructing a candidate box regression network, the pyramid technology adopted in previous inventions is not needed, and a lot of computing resources are saved.

Description

technical field [0001] The invention belongs to the technical field of pattern recognition, in particular to an end-to-end convolutional neural network pedestrian detection method. Background technique [0002] Pattern recognition is an important interdisciplinary subject in the field of artificial intelligence and image processing, and it is a research hotspot in recent years. Early task solutions mainly relied on human professional domain knowledge to design an algorithm or build a system, and the problem-solving effect was limited by human knowledge. With the development of artificial intelligence, the convolutional neural network was formally proposed by Lecun in the 1990s. Hinton improved the original convolutional neural network in 2012 and achieved the first result in the ImageNet competition. Since then, convolutional neural networks have been widely used in fields such as computer vision, natural language processing, and intelligent search. Convolutional neural n...

AI Technical Summary

Problems solved by technology

[0003] After searching the existing technology, it was found that the Chinese patent document number CN105335716A was published (announced) on 2016.02.17, which disclosed a pedestrian detection method based on improved UDN to extract joint features, including: image preprocessing; convolutional neural network based on The preprocessed image extracts the overall and local features of the human body; weights the category probabilities output by the overall features and local features of the step to obtain the final probability value, and judges whether the original input image contains pedestrians according to the final probability value. Compared with the low missed detection rate, the actual missed detection rate is still high, and the technology has not achieved end-to-end training, and must rely on the HOG+CSS++SVM algorithm to obtain the initial candidate area for pedestrians

This technology requires module cascading, which makes the whole system more complicated

Images