loss layer of implementation.
You can see "Focal Loss for Dense Object Detection" arXiv for more information.
// Focal Loss layer
optional FocalLossParameter focal_loss_param = 124;
// Focal Loss for Dense Object Detection
message FocalLossParameter {
enum Type {
ORIGIN = 0; // FL(p_t) = -(1 - p_t) ^ gama * log(p_t), where p_t = p if y == 1 else 1 - p, whre p = sigmoid(x)
LINEAR = 1; // FL*(p_t) = -log(p_t) / gama, where p_t = sigmoid(gama * x_t + beta), where x_t = x * y, y is the ground truth label {-1, 1}
}
optional Type type = 1 [default = ORIGIN];
optional float gamma = 2 [default = 2];
// cross-categories weights to solve the imbalance problem
optional float alpha = 3 [default = 0.25];
optional float beta = 4 [default = 1.0];
}
layer {
name: "loss_cls"
type: "FocalLoss"
bottom: "cls_score"
bottom: "labels"
propagate_down: 1
propagate_down: 0
top: "loss_cls"
include { phase: TRAIN }
loss_weight: 1
loss_param { ignore_label: -1 normalize: true }
focal_loss_param { alpha: 0.25 gamma: 2 }
}
see https://github.com/zimenglan-sysu-512/paper-note/blob/master/focal_loss.pdf
All categories share the same alpha.
Here use softmax instead of sigmoid function.
If you want see how to use sigmoid to implement Focal Loss, please see https://github.com/sciencefans/Focal-Loss to get more information.