Vehicle Make and Model Recognition in CCTV footage

Part‐based recognition of vehicle make and model

“…After ROI detection, a simplified version of SIFT and LDA have been used for feature extraction and k‐NN as the classifier. Some other researchers used similar approaches for ROI localisation [12–14].…”

Car detection and classification using cascade model

“…Recent few years, many models such as deformable part model [4–7], constellation model [7–9], graphical model [10], amount of part‐based models [4–6, 8, 11–13], and some kinds of deformable template models [14–16] have been proposed to figure out the object problems. To overcome the issues of vehicle‐type recognition, the lectures [17–21] had put forward some feasible methods. Since the car's front face contains a large amount of information such as plates, logo, shape of headlights etc.…”

“…Since the car's front face contains a large amount of information such as plates, logo, shape of headlights etc. The lectures [17, 18] have made a preliminary attempt to use the car's front face to identify the vehicle type. Zhang et al [20] used histograms of oriented gradients features of a vehicle frontal face to improve the recognition rate.…”

Fusion with CS-LBP and HOG for Vehicle Make and Model Recognition

“…These logo, grid and headlights constitute the face characteristics of vehicles which provide evidence to the identification of vehicles. In the paper [5,6,7,8], the vehicle's face image is attempted to identify the specific vehicle models, for example, paper [7] used the shape histogram features of the vehicle's face image and the Bayesian transcendental model to identify the specific vehicle model, paper [8] used the HOG feature which is extracted from the face image of the vehicle to identify 80 vehicle models of 15 brands. The above methods verify the feasibility of identifying the specific vehicle models based on the vehicle's face images, but the recognition rate needs to be improved.…”