Deepsort-based multi-object face tracking method and apparatus, computer device, and medium

A computer device, comprising one or more processors and a machine-readable storage medium, the machine-readable storage medium storing machine-executable instructions executable by the one or more processors, wherein the machine-executable instructions, when executed by the one or more processors, cause the one or more processors to perform operations comprising:

• performing object detection on a current video frame to obtain a face detection box set and an upper body detection box set;
• for an upper body detection box of the upper body detection box set, determining, by a deep simple online and realtime tracking (DeepSORT) algorithm, a trajectory corresponding to the upper body detection box and a predicted upper body detection box matching the upper body detection box, wherein an object identity (ID) is bound to the trajectory;
• and
• determining, based on a positional relationship between each face detection box of the face detection box set and a designated region in a designated upper body detection box, a face detection box corresponding to the upper body detection box, wherein the designated upper body detection box is an upper body detection box of the upper body detection box set or the predicted upper body detection box, and an ID of the face detection box is the object ID bound to the trajectory corresponding to the face detection box.

A non-transitory computer-readable storage medium, storing computer programs thereon, wherein the computer programs, when executed by one or more processors, cause the one or more processors to perform operations comprising:

• performing object detection on a current video frame to obtain a face detection box set and an upper body detection box set;
• for an upper body detection box of the upper body detection box set, determining, by a deep simple online and realtime tracking (DeepSORT) algorithm, a trajectory corresponding to the upper body detection box and a predicted upper body detection box matching the upper body detection box, wherein an object identity (ID) is bound to the trajectory;
• and
• determining, based on a positional relationship between each face detection box of the face detection box set and a designated region in a designated upper body detection box, a face detection box corresponding to the upper body detection box, wherein the designated upper body detection box is an upper body detection box of the upper body detection box set or the predicted upper body detection box, and an ID of the face detection box is the object ID bound to the trajectory corresponding to the face detection box.