Key Frame Extraction for Sports Training Based on Improved Deep Learning

<table class="table-group" id="tab1"><tr><td><table class="table"><tr><td class="thead-hr" colspan="2"><hr/></td></tr><tr class="thead"><td class="align_left">Operation</td><td class="align_center">Description</td></tr><tr><td class="thead-hr" colspan="2"><hr/></td></tr><tr><td class="align_left">Botha’</td><td class="align_center">It is a morphological “bottom cap” transformation operation, and the returned image is the original image minus the morphological closing operation (closing operation: first expand and then corrode)</td></tr><tr><td class="align_left">Bridge</td><td class="align_center">Disconnected pixels: the value pixel is set to 1 if it has two nonzero unconnected (8 neighborhood) pixels</td></tr><tr><td class="align_left">Clean</td><td class="align_center">Remove isolated pixels (by О 1)</td></tr><tr><td class="align_left">Close</td><td class="align_center">Perform morphological closing operation (expansion before corrosion)</td></tr><tr><td class="align_left">Diag</td><td class="align_center">The diagonal filling is used to eliminate the 8 connected regions in the background</td></tr><tr><td class="align_left">Dilate</td><td class="align_center">The structure ones (3) are used to perform the expansion operation</td></tr><tr><td class="align_left">Erode</td><td class="align_center">The structure ones (3) are used to perform the corrosion operation</td></tr><tr><td class="align_left">Fill</td><td class="align_center">Fill in isolated internal pixels (0 surrounded by 1)</td></tr><tr class="table-tr"><td colspan="2"><hr class="tbody-hr"/></td></tr></table></td></tr></table>

<div>bwmorph morphological operations on images.</div>

Scientific Programming

tab1

Table 1

Table 1: Key Frame Extraction for Sports Training Based on Improved Deep Learning 