loss_functions.py

import tensorflow as tf
from keras import backend as K


def _to_tensor(x, dtype):
    """Convert the input `x` to a tensor of type `dtype`.
    # Arguments
    x: An object to be converted (numpy array, list, tensors).
    dtype: The destination type.
    # Returns
    A tensor.
    """
    x = tf.convert_to_tensor(x)
    if x.dtype != dtype:
        x = tf.cast(x, dtype)
    return x
    
    
def cross_entropy_loss_RCF(y_true, y_pred):
    _epsilon = _to_tensor(K.epsilon(), y_pred.dtype.base_dtype)
    y_pred = tf.clip_by_value(y_pred, _epsilon, 1 - _epsilon)
    y_pred = tf.log(y_pred / (1 - y_pred))

    y_true = tf.cast(y_true, tf.float32)
    lamda = 1.2
    count_neg = tf.reduce_sum(1. - y_true)
    count_pos = tf.reduce_sum(y_true)
    alpha = lamda * count_pos / (count_neg + count_pos)
    beta = count_neg / (count_neg + count_pos)

    pos_weight = beta / alpha

    cost = tf.nn.weighted_cross_entropy_with_logits(logits=y_pred, targets=y_true, pos_weight=pos_weight)

    # Multiply by 1 - beta
    cost = tf.reduce_mean(cost * (1 - beta))

    # check if image has no edge pixels return 0 else return complete error function
    return tf.where(tf.equal(count_pos, 0.0), 0.0, cost)

def cross_entropy_balanced(y_true, y_pred):
    """
    Implements Equation [2] in https://arxiv.org/pdf/1504.06375.pdf
    Compute edge pixels for each training sample and set as pos_weights to tf.nn.weighted_cross_entropy_with_logits
    """
    # Note: tf.nn.sigmoid_cross_entropy_with_logits expects y_pred is logits, Keras expects probabilities.
    # transform y_pred back to logits
    _epsilon = _to_tensor(K.epsilon(), y_pred.dtype.base_dtype)
    y_pred = tf.clip_by_value(y_pred, _epsilon, 1 - _epsilon)
    y_pred = tf.log(y_pred / (1 - y_pred))

    y_true = tf.cast(y_true, tf.float32)

    count_neg = tf.reduce_sum(1. - y_true)
    count_pos = tf.reduce_sum(y_true)

    # Equation [2]
    beta = count_neg / (count_neg + count_pos)

    # Equation [2] divide by 1 - beta
    pos_weight = beta / (1 - beta)

    cost = tf.nn.weighted_cross_entropy_with_logits(logits=y_pred, targets=y_true, pos_weight=pos_weight)

    # Multiply by 1 - beta
    cost = tf.reduce_mean(cost * (1 - beta))

    # check if image has no edge pixels return 0 else return complete error function
    return tf.where(tf.equal(count_pos, 0.0), 0.0, cost)


def pixel_error(y_true, y_pred):

    pred = tf.cast(tf.greater(y_pred, 0.5), tf.int32, name='predictions')
    error = tf.cast(tf.not_equal(pred, tf.cast(y_true, tf.int32)), tf.float32)
    return tf.reduce_mean(error, name='pixel_error')