Method and apparatus for training target detection model, method and apparatus for detecting target

A computer device, comprising:

• a processor and a memory;
• wherein
• the memory is configured to store a computer program;
• and
• the processor, when running the computer program stored in the memory, is caused to perform:
• determining a first region in a sample image, wherein the first region is a target region predicted by the target detection model in the sample image, and the sample image comprises one or more of fireworks, dust, clouds, shorelines, flame, smoke, steam, or night light;
• counting training periods of the target detection model and determining a relationship between an intersection region and the first region in response to a number of training periods of the target detection model reaching a target number, wherein the intersection region is an intersection of the first region and a second region, wherein the second region is a region annotated for the target in the sample image in a data annotation phase, and surrounds the target in the sample image;
• determining, in response to the relationship between the intersection region and the first region satisfying a target relationship, a predetermined low loss function value as a loss function value of the first region, wherein the predetermined low loss function value is a constant and is less than 0.001;
• and
• training the target detection model with reference to the loss function value;
• wherein the target detection model is used to detect one or more of fireworks, dust, clouds, shorelines, flame, smoke, steam, or night light in an image;
• wherein determining the relationship between the intersection region and the first region comprises:
• calculating a ratio of an area of the intersection region to an area of the first region;
• the relationship between the intersection region and the first region satisfying the target relationship comprises:
• the ratio of the area of the intersection region to the area of the first region being not less than a threshold, the threshold ranging from 0.9 to 1;
• and
• training the target detection model with reference to the loss function value comprises:
• updating a parameter of the target detection model with reference to the loss function value that is suppressed, and performing a next period of training.

A non-transitory computer-readable storage medium, storing a computer instruction therein, wherein the stored computer instruction, when executed by a processor, causes the processor to perform:

• determining a first region in a sample image, wherein the first region is a target region predicted by the target detection model in the sample image, and the sample image comprises one or more of fireworks, dust, clouds, shorelines, flame, smoke, steam, or night light;
• counting training periods of the target detection model and determining a relationship between an intersection region and the first region in response to a number of training periods of the target detection model reaching a target number, wherein the intersection region is an intersection of the first region and a second region, wherein the second region is a region annotated for the target in the sample image in a data annotation phase, and surrounds the target in the sample image;
• determining, in response to the relationship between the intersection region and the first region satisfying a target relationship, a predetermined low loss function value as a loss function value of the first region, wherein the predetermined low loss function value is a constant and is less than 0.001;
• and
• training the target detection model with reference to the loss function value;
• wherein the target detection model is used to detect one or more of fireworks, dust, clouds, shorelines, flame, smoke, steam, or night light in an image;
• wherein determining the relationship between the intersection region and the first region comprises:
• calculating a ratio of an area of the intersection region to an area of the first region;
• the relationship between the intersection region and the first region satisfying the target relationship comprises:
• the ratio of the area of the intersection region to the area of the first region being not less than a threshold, the threshold ranging from 0.9 to 1;
• and
• training the target detection model with reference to the loss function value comprises:
• updating a parameter of the target detection model with reference to the loss function value that is suppressed, and performing a next period of training.