Machine learning model training method and device, and expression image classification method and device

A computing device comprising one or more processors, memory coupled to the one or more processors and a plurality of programs stored in the memory that, when executed by the one or more processors, cause the computing device to perform a plurality of operations comprising:

• obtaining a machine learning model through training according to a general-purpose image training set, wherein the machine learning model comprises one or more model parameters;
• determining a sample set of special-purpose images and a corresponding classification label, further including:
• obtaining a representative special-purpose image corresponding to the classification label and an image feature of the representative special-purpose image from an intermediate layer in the machine learning model;
• and
• determining, according to similarities between the image feature of the representative special-purpose image and corresponding image features of the special-purpose images in a special-purpose image library generated by the intermediate layer in the machine learning model, a set of special-purpose images whose similarity to the representative special-purpose image exceeds a predefined threshold as the sample set of special-purpose images corresponding to the classification label;
• inputting the sample set of special-purpose images to the machine learning model, to obtain an intermediate classification result for the corresponding classification label;
• adjusting the model parameters of the machine learning model according to a difference between the intermediate classification result and the classification label;
• and
• repeating the inputting and adjusting steps until that a training stop condition is met, to obtain the machine learning model with the adjusted model parameters.

A non-transitory computer readable storage medium storing a plurality of machine readable instructions in connection with a computing device having one or more processors, wherein the plurality of machine readable instructions, when executed by the one or more processors, cause the computing device to perform a plurality of operations including:

• obtaining a machine learning model through training according to a general-purpose image training set, wherein the machine learning model comprises one or more model parameters;
• determining a sample set of special-purpose images and a corresponding classification label, further including:
• obtaining a representative special-purpose image corresponding to the classification label and an image feature of the representative special-purpose image from an intermediate layer in the machine learning model;
• and
• determining, according to similarities between the image feature of the representative special-purpose image and corresponding image features of the special-purpose images in a special-purpose image library generated by the intermediate layer in the machine learning model, a set of special-purpose images whose similarity to the representative special-purpose image exceeds a predefined threshold as the sample set of special-purpose images corresponding to the classification label;
• inputting the sample set of special-purpose images to the machine learning model, to obtain an intermediate classification result for the corresponding classification label;
• adjusting the model parameters of the machine learning model according to a difference between the intermediate classification result and the classification label;
• and
• repeating the inputting and adjusting steps until that a training stop condition is met, to obtain the machine learning model with the adjusted model parameters.