GP22: A Car Styling Dataset for Automotive Designers

GP22 is the dataset relevant to the designer's perspective and practices. The dataset involves side profiles of cars produced or introduced from the 1960s to 2021. Images folder contains dataset images labeled with: SEGEMENT_BRAND_COUNT.jpg SEGMENT were labeled in single alphabet: A, B, C, D, E, F, J, S, P BRAND were labeled in three alphabet: Alf(AlfaRomeo), Ast(Aston Martin), Aud(Audi), Ben(Bentley), Bmw(BMW), Che(Chevrolet), Chr(Chrysler), Cit(Citroen), Fer(Ferrari), Fia(Fiat), For(Ford), Gen(Genesis), GMC(GMC), Hon(Honda), Hum(Hummer), Hyu(Hyundai), Jag(Jaguar), Jep(Jeep), Kia(Kia), Lan(LandRover), Lex(Lexus), Mas(Maserati), Maz(Mazda), Mer(Mercedes Benz), Mit(Mitsubishi), Nis(Nissan), Ope(Opel), Peu(Peugeot), Pon(Pontiac), Por(Porsche), Ren(Renault), Riv(Rivian), Rol(Rolls Royce), Sko(Skoda), Ssa(Ssang Yong), Vwg(Volkswagen), and Vol(Volvo) Labels folder contain labels in text file: SEGEMENT_BRAND_COUNT.txt Single text file structured in: 0(tire1): x1, y1, x2, y2 positions of the single box 0(tire2): x1, y1, x2, y2 positions of the single box 1(wheel1): x1, y1, x2, y2 positions of the single box 1(wheel2): x1, y1, x2, y2 positions of the single box 2(body): x1, y1, x2, y2 positions of the single box 3(cabin): x1, y1, x2, y2 positions of the single box 4(day light opening): x1, y1, x2, y2 positions of the single box 5(bodyside): x1, y1, x2, y2 positions of the single box This dataset also contained the YOLO v5 l based detection model trained on GP22: DetectionModel.pt