Image Recognition
In this article, we discuss semantic segmentation using TensorFlow Keras. Among FCNNs, U-Net is one of the successful architectures acclaimed for its performance in Medical Image Segmentation. def brightness(img, mask): img = tf.image.adjustbrightness(img, 0.1) return img, mask def gamma(img, mask): img = tf.image.adjustgamma(img, 0.1) return img, mask def hue(img, mask): img = tf.image.adjusthue(img, -0.1) return img, mask def crop(img, mask): img = tf.image.centralcrop(img, 0.7) img = tf.image.resize(img, (128,128)) mask = tf.image.centralcrop(mask, 0.7) mask = tf.image.resize(mask, (128,128)) mask = tf.cast(mask, tf.uint8) return img, mask def fliphori(img, mask): img = tf.image.flipleftright(img) mask = tf.image.flipleftright(mask) return img, mask def flipvert(img, mask): img = tf.image.flipupdown(img) mask = tf.image.flipupdown(mask) return img, mask def rotate(img, mask): img = tf.image.rot90(img) mask = tf.image.rot90(mask) return img, maskApply augmentation to the data with the above functions. train = tf.data.Dataset.zip((trainX, trainy)) val = tf.data.Dataset.zip((valX, valy)) # perform augmentation on train data only a = train.map(brightness) b = train.map(gamma) c = train.map(hue) d = train.map(crop) e = train.map(fliphori) f = train.map(flip_vert) g = train.map(rotate) train = train.concatenate(a) train = train.concatenate(b) train = train.concatenate(c) train = train.concatenate(d) train = train.concatenate(e) train = train.concatenate(f) train = train.concatenate(g)Prepare data batches. Wrapping UpIn this article, we have discussed semantic segmentation using TensorFlow Keras.