Ensemble machine learning model architecture for lesion detection

A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device, causes the computing device to implement a lesion detection ensemble machine learning model architecture comprising a plurality of trained machine learning computer models, wherein the lesion detection ensemble machine learning model architecture executes on the computing device to:

• process, by a first decoder of a lesion detection machine learning computer model, a medical image input of at least one medical image to generate a first lesion mapping prediction output identifying graphical elements corresponding to lesions in the at least one medical image;
• process, by a second decoder of the lesion detection machine learning computer model, the medical image input to generate a second lesion mapping prediction output identifying graphical elements corresponding to lesions in the at least one medical image;
• combine, by combinational logic of the lesion detection machine learning computer model, the first lesion mapping prediction output and second lesion mapping prediction output to generate a combined lesion mapping prediction output;
• generate, by final lesion mapping output logic of the lesion detection ensemble machine learning model architecture, a final lesion prediction output based on the combined lesion mapping prediction output;
• and
• output, by the final lesion mapping output logic, the final lesion prediction output for further downstream computing operations, wherein the first decoder is trained with a first loss function that is configured to counterbalance a training of the second decoder that is trained using a second loss function different from the first loss function.

An apparatus comprising:

• at least one processor;
• and
• at least one memory coupled to the at least one processor, wherein the at least one memory comprises instructions which, when executed by the at least one processor, cause the at least one processor to implement a lesion detection ensemble machine learning model architecture comprising a plurality of trained machine learning computer models, wherein the lesion detection ensemble machine learning model architecture executes on the apparatus to:
• process, by a first decoder of a lesion detection machine learning computer model, a medical image input of at least one medical image to generate a first lesion mapping prediction output identifying graphical elements corresponding to lesions in the at least one medical image;
• process, by a second decoder of the lesion detection machine learning computer model, the medical image input to generate a second lesion mapping prediction output identifying graphical elements corresponding to lesions in the at least one medical image;
• combine, by combinational logic of the lesion detection machine learning computer model, the first lesion mapping prediction output and second lesion mapping prediction output to generate a combined lesion mapping prediction output;
• generate, by final lesion mapping output logic of the lesion detection ensemble machine learning model architecture, a final lesion prediction output based on the combined lesion mapping prediction output;
• and
• output, by the final lesion mapping output logic, the final lesion prediction output for further downstream computing operations, wherein the first decoder is trained with a first loss function that is configured to counterbalance a training of the second decoder that is trained using a second loss function different from the first loss function.