Person Retrieval in Surveillance Video using Height, Color and Gender

Visual appearance based person retrieval in unconstrained environment videos

“…Deep learning based methodologies are widely used for semantic description based person retrieval due to its efficient learning. Convolutional Neural Network (CNN) based algorithms are reported for person attribute recognition [10][11][12] as well as for person retrieval [5,[13][14][15][16]. Semantic Retrieval Convolution Neural Network (SRCNN) [10] retrieves the soft biometrics with recognition rate of 20.1% and 46.4% at rank-1 for one-shot and multi-shot identification respectively on SoBiR [8] dataset.…”

“…The outcome of the AVSS challenge II showcased that CNN based algo rithms achieved higher accuracy. Galiyawala et al [15] propose a linear filtering approach which reduces the search space during the retrieval. Semantic segmentation nullifies limitations of the bounding box based approaches.…”

“…The algorithm achieves an average Intersection-over-Union (IoU) of 0.36 and 52.2% of frames have an IoU greater than 0.4. The baseline approach [7] and [15] uses subset of available soft biometrics, limiting their accuracy. The color patch extraction is also noisy in [15].…”

“…The baseline approach [7] and [15] uses subset of available soft biometrics, limiting their accuracy. The color patch extraction is also noisy in [15]. Moreover, [15] uses AlexNet [25] which has lower retrieval accuracy compared to deeper networks like ResNet [26] and DenseNet [27].…”

“…The color patch extraction is also noisy in [15]. Moreover, [15] uses AlexNet [25] which has lower retrieval accuracy compared to deeper networks like ResNet [26] and DenseNet [27]. The transfer learning based approach by Yaguchi et al [14] also uses Mask R-CNN for person detection and DenseNet-161 for soft biometric classification.…”