Deep convolution neural network with context based expanded neighbourhoods distance re-ranking model for person re-identification

Presently, Person Re-Identification (PRe-ID) acts as a vital part of real time video surveillance to ensure the rising need for public safety. Resolving the PRe-ID problem includes the process of matching observations of persons among distinct camera views. Earlier models consider PRe-ID as a unique object retrieval issue and determine the retrieval results mainly based on the unidirectional matching among the probe and gallery images. But the accurate matching might not present in the top-k ranking results owing to the appearance modifications caused by the difference in illumination, pose, viewpoinst, and occlusion. For addressing these issues, this paper presents new deep learning (DL) with expanded neighbourhood distance reranking (DL-ENDR) model for PRe-ID. The proposed DL-ENDR involves different processes for PRe-ID, such as feature extraction, similarity measurement, and feature re-ranking. The DL-ENDR model uses a Densely Connected Convolutional Networks (DenseNet169) model as a feature extractor. Additionally, Euclidean distance-based similarity measurement is employed to determine the resemblance between the probe and gallery images. Finally, the DL-ENDR model incorporated ENDR model to re-rank the outcome of the person-reidentification along with Mahala Nobis distance. An extensive experimental analysis takes place on benchmark dataset and the obtained results verified the effective performance of the DL-ENDR model interms of different aspects.

Automated summary

This paper presents a new deep learning model with expanded neighbourhood distance reranking (DL-ENDR) for Person Re-Identification (PRe-ID). The DL-ENDR model addresses the challenges in PRe-ID through processes like feature extraction, similarity measurement, and feature re-ranking. Experimental analysis on benchmark dataset confirms the effective performance of the DL-ENDR model in various aspects.