Zhao Qingli

TL;DR: This paper proposes multi-label-based similarity learning (MLSL) for vehicle re-identification obtaining an efficient deep-learning-based model that derives robust vehicle representations and proves the superiority of the model over multiple state-of-the-art methods on the three mentioned datasets.

TL;DR: This work proposed a deep learning based framework that can lead to an efficient representation of vehicles and variational feature learning is employed to generate variational features which are more discriminating.

TL;DR: Li et al. as discussed by the authors proposed a target detection method aiming at a complex scene, comprising the following steps: building an SSD target detection model, which consists a front basic network connected in turn, a feature extraction network and a classification detection module; modifying the front-end basic network and the first two feature extraction layers of the SSD model, adding and modifying Inception component to improve the multi-scale adaptability of SSD model.

TL;DR: In this paper, a large-angle license plate inclination correction method based on an end-to-end neural network is proposed, which comprises the following steps: obtaining a rough license plate area by using a license plate detection technology, and expanding the license plate image to obtain an expanded license plate as an input area; extracting high-dimensional features from the input area by extracting a feature map; analyzing the feature map to obtain absolute coordinates offour vertexes of the candidate license plate areas; and correcting the inclination of the license plates according to the absolute coordinates of the four vertexes

TL;DR: In this article, a low-resolution pedestrian detection method and system and a storage medium is proposed, which comprises a training process and a testing process, wherein the training process firstly determines parameters of a training set, and then sequentially inputting the pictures according to the batch processing size, extracting multi-scale features of the training pictures, and reconstructing and enhancing shallow features to form a new multiscale detection framework; and finally, carrying out frame classification and position regression.