Computer Vision News - June 2020

2 Summary Co puter Vision Tools 36 Here’s the pixel-wise segmentation! I hope you enjoyed this article. See you next month and till then, stay safe! plt.imshow(final_output) plt.figure() plt.imshow(x[PATCH_SIZE[0] : x.shape[0]-PATCH_SIZE[0], PATCH_SIZE[1] : x.shape[1]-PATCH_SIZE[1]]) out_image = np.array(out_image) out_image = np.exp(out_image) # log softmax output from torch. Nullify log by applying exp operation final_output = np.argmax(out_image, axis=2).reshape(128, 128) # argmax along the probability dimension Now the final output can be plotted: