Robust face detection

A device, comprising:

• a camera;
• an illuminator;
• a memory configured to store program instructions;
• and
• a processor configured to execute the program instructions and to cause the device to:
• capture at least one image using the camera, wherein the at least one image comprises an image captured while illuminating subjects in the image with the illuminator;
• encode the at least one image to generate feature vectors defining features of subjects in the image in a feature space;
• assess the generated feature vectors in the feature space to detect a face in the image, wherein at least some part of the face is present in the image, and wherein the face is detected regardless of an orientation of the face with respect to the camera;
• form a bounding box around the face detected in the image, wherein, if a portion of the face detected is located outside an image boundary, the bounding box is formed around the at least some part of the face detected to be present in the image;
• and
• provide an output comprising the bounding box formed for the detected face in the image.

A method, comprising:

• capturing at least one image using a camera located on a device, the device comprising a computer processor and a memory, wherein the at least one image comprises an image captured while illuminating subjects in the image with an illuminator located on the device;
• encoding the at least one image to generate feature vectors defining features of subjects in the image in a feature space, wherein an exposure in the at least one image is in an exposure range set by exposure limits of the camera on the device;
• assessing the generated feature vectors in the feature space to detect a face in the image, wherein at least some part of the face is present in the image;
• forming a bounding box around the face detected in the image;
• and
• providing an output, the output comprising the bounding box formed for the detected face in the image.