nodes/set-soft-ComfyUI-ImageMisc
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[SaliencyEvaluationMetrics][Added] Conditional computation

Browse Source
This commit is contained in:
Salvador E. Tropea
2025-11-06 20:43:06 -03:00
parent 4d03617fc6
commit a2ca2fe876
1 changed files with 67 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/nodes/nodes_img.py
+67 -32
View File
@@ -770,16 +770,36 @@ class SaliencyEvaluationMetrics:
"prediction": ("MASK",),
"ground_truth": ("MASK",),
},
"optional": {
"mae_enable": ("BOOLEAN", {
"default": True,
"tooltip": "Compute the MAE"}),
"max_f_mes_enable": ("BOOLEAN", {
"default": True,
"tooltip": "Compute the Max_F-measure"}),
"s_mes_enable": ("BOOLEAN", {
"default": True,
"tooltip": "Compute the S-measure"}),
"e_mes_enable": ("BOOLEAN", {
"default": True,
"tooltip": "Compute the E-measure"}),
"wf_mes_enable": ("BOOLEAN", {
"default": True,
"tooltip": "Compute the Weighted F-measure"}),
},
}
RETURN_TYPES = ("FLOAT", "FLOAT", "FLOAT", "FLOAT", "FLOAT")
RETURN_NAMES = ("MAE", "Max_F-measure", "S-measure", "E-measure", "Weighted_F-measure")
RETURN_TYPES = ("DICT", "FLOAT", "FLOAT", "FLOAT", "FLOAT", "FLOAT")
RETURN_NAMES = ("all", "MAE", "Max_F-measure", "S-measure", "E-measure", "Weighted_F-measure")
OUTPUT_IS_LIST = (True, False, False, False, False, False)
FUNCTION = "evaluate"
CATEGORY = BASE_CATEGORY + "/" + "Analysis"
UNIQUE_NAME = "SET_SaliencyEvaluationMetrics"
DISPLAY_NAME = "Saliency Evaluation Metrics"
def evaluate(self, prediction: torch.Tensor, ground_truth: torch.Tensor):
def evaluate(self, prediction: torch.Tensor, ground_truth: torch.Tensor, mae_enable: bool = True,
max_f_mes_enable: bool = True, s_mes_enable: bool = True, e_mes_enable: bool = True,
wf_mes_enable: bool = True):
# Ensure tensors are on the same device and float type
device = get_default_comfy_device()
inputs_are_copies = get_canonical_device(prediction.device) != device
@@ -799,54 +819,69 @@ class SaliencyEvaluationMetrics:
mae_total, f_measure_max_total, s_measure_total, e_measure_total, weighted_f_total = 0, 0, 0, 0, 0
eps = 1e-6
all = []
for i in range(batch_size):
pred_i = pred[i]
gt_i = gt[i]
res = {}
# 1. Mean Absolute Error (MAE)
mae = torch.mean(torch.abs(pred_i - gt_i))
logger.debug(f"MAE: {mae}")
mae_total += mae
if mae_enable:
mae = torch.mean(torch.abs(pred_i - gt_i))
logger.debug(f"MAE: {mae}")
mae_total += mae
res['mae'] = mae
# --- Metrics requiring binary ground truth ---
gt_binary = (gt_i >= 0.5).float()
if max_f_mes_enable or s_mes_enable or e_mes_enable or wf_mes_enable:
gt_binary = (gt_i >= 0.5).float()
# 2. Max F-measure
f_max = 0.0
for threshold in torch.linspace(0, 1, 256, device=device):
pred_binary = (pred_i >= threshold).float()
if max_f_mes_enable:
f_max = 0.0
for threshold in torch.linspace(0, 1, 256, device=device):
pred_binary = (pred_i >= threshold).float()
tp = (pred_binary * gt_binary).sum()
tp = (pred_binary * gt_binary).sum()
if tp == 0:
continue
if tp == 0:
continue
precision = tp / (pred_binary.sum() + eps)
recall = tp / (gt_binary.sum() + eps)
precision = tp / (pred_binary.sum() + eps)
recall = tp / (gt_binary.sum() + eps)
# Using beta^2 = 0.3 as is standard.
beta2 = 0.3
f_beta = (1 + beta2) * precision * recall / (beta2 * precision + recall + eps)
# Using beta^2 = 0.3 as is standard.
beta2 = 0.3
f_beta = (1 + beta2) * precision * recall / (beta2 * precision + recall + eps)
if f_beta > f_max:
f_max = f_beta
f_measure_max_total += f_max
logger.debug(f"F_max: {f_max}")
if f_beta > f_max:
f_max = f_beta
f_measure_max_total += f_max
logger.debug(f"F_max: {f_max}")
res['max_f_mes'] = f_max
# 3. S-measure
s_measure = _get_s_measure(pred_i, gt_binary)
s_measure_total += s_measure
logger.debug(f"S: {s_measure}")
if s_mes_enable:
s_measure = _get_s_measure(pred_i, gt_binary)
s_measure_total += s_measure
logger.debug(f"S: {s_measure}")
res['s_mes'] = s_measure
# 4. E-measure
e_measure = _get_e_measure(pred_i, gt_binary)
e_measure_total += e_measure
logger.debug(f"E: {e_measure}")
if e_mes_enable:
e_measure = _get_e_measure(pred_i, gt_binary)
e_measure_total += e_measure
logger.debug(f"E: {e_measure}")
res['e_mes'] = e_measure
# 5. Weighted F-measure
wf = _get_weighted_f_measure(pred_i, gt_binary)
weighted_f_total += wf
logger.debug(f"wF: {wf}")
if wf_mes_enable:
wf = _get_weighted_f_measure(pred_i, gt_binary)
weighted_f_total += wf
logger.debug(f"wF: {wf}")
res['wf_mes'] = wf
all.append(res)
# --- Average metrics over the batch ---
mae_avg = mae_total.item() / batch_size
@@ -855,7 +890,7 @@ class SaliencyEvaluationMetrics:
e_measure_avg = e_measure_total.item() / batch_size
weighted_f_avg = weighted_f_total.item() / batch_size
return (mae_avg, f_measure_avg, s_measure_avg, e_measure_avg, weighted_f_avg)
return (all, mae_avg, f_measure_avg, s_measure_avg, e_measure_avg, weighted_f_avg)
class CompositeFace: