Transductive Transfer Learning to Specialize a Generic Classifier Towards a Specific Scene

Abstract

In this paper, we tackle the problem of domain adaptation to perform object-classification and detection tasks in video surveillance starting by a generic trained detector. Precisely, we put forward a new transductive transfer learning framework based on a sequential Monte Carlo filter to specialize a generic classifier towards a specific scene. The proposed algorithm approximates iteratively the target distribution as a set of samples (selected from both source and target domains) which feed the learning step of a specialized classifier. The output classifier is applied to pedestrian detection into a traffic scene. We have demonstrated by many experiments, on the CUHK Square Dataset and the MIT Traffic Dataset, that the performance of the specialized classifier outperforms the generic classifier and that the suggested algorithm presents encouraging results.

Sequential Monte Carlo filter based on multiple strategies for a scene specialization classifier

“…It can be used to specialize any classifier while utilizing the same function as a generic classifier and may be applied using several strategies on each step of the filter. Some preliminary results of the work presented in this paper were published in [20]. In this paper, we put forward an extension of our original TTL approach based on an SMC (TTL-SMC) filter by other sample proposal and observation strategies and more experiments.…”

“…Mainly three categories of transfer learning methods, related to the suggested approach, were described in [20]. The first category would modify the parameters of a source learning model to improve its accuracy in a target domain [30,31].…”

“…Nevertheless, it was demonstrated in [15][16][17][18][19][20] that the accuracy of a generic (pedestrian or vehicle) detector would drop-off quickly when it was applied to a specific traffic scene, in which the available data would mismatch the training source one.…”

SMC faster R-CNN: Toward a scene-specialized multi-object detector

“…In the recent years, transfer learning has attracted a lot of research groups in developing state-of-the-art theories and new applications in several domains like object detection and recognition [5][6][7] [11]. Transfer learning aims to address the problem when the distribution of the training data from the source domain is different from that of the target one.…”

“…The second category would decrease the variation between the source and target distributions to adapt a detector to the target domain [14] [15]. The third one would automatically choose the training samples that could provide a better detector or classifier for the target task [5] [11]. In this paper, we focus on the third category which utilizes an automatic labeler to select data from the target domain.…”