entry_exit_mouse_box.measures module

class entry_exit_mouse_box.measures.MiceVisibilityProcessor(mask_path, areas, ma, start, duration)

Bases: object

Calculates an array indicating for each box (designated by the labels in ‘areas’) if a mouse is inside or not. The process is realized on several threads. The input video was saved as a mask (0=BG, 255=FG) but it requires thresholding anyway due to compression. The areas are a grayscale image with one value per box (0=BG). To process the presence of a mouse, we use the length of the ellipse fitted to the mouse’s label. It requires the input image to be calibrated. The process doesn’t start from the frame 0 but from the frame ‘start’. We don’t need a control structure to write in the buffer as the threads are not writing in the same place.

fix_visibility()

process_n_in_out()

process_sessions()

We call ‘session’ the span of time during which the mouse is hidden or visible. For each box, a video is a succession of sessions, alternating between hidden and visible. During a session, the mouse is either hidden or visible. A session is defined by a duration (in seconds) and a distance (in pixels).

process_visibility_pos(batch, frames)

read_frames(frames_count=32)

release_resources()

smooth_centroids()

start_processing(num_workers=4)

worker()

class entry_exit_mouse_box.measures.QtWorkerMVP(mask_path, areas, ma, start, duration)

Bases: QObject

measures_ready

run()

entry_exit_mouse_box.measures.calculate_maximal_length(mask)