File size: 5,064 Bytes
afabda4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
# -*- coding: utf-8 -*-
"""

Created on Mon Jun 30 17:06:08 2025



@author: User

"""

import torch
import torch.nn as nn
import numpy as np
import librosa
import joblib
import pickle
from pathlib import Path
from sklearn.isotonic import IsotonicRegression
import argparse

# ==== CONFIGURACIÓN ====
SR = 22050
DURATION = 4.0
SAMPLES = int(SR * DURATION)
BANDS = 128
HOP = 512
FMIN, FMAX = 150, 4500
DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")

# ==== MODELO ====
class SEBlock(nn.Module):
    def __init__(self, channels, red=16):
        super().__init__()
        self.fc = nn.Sequential(
            nn.AdaptiveAvgPool2d(1),
            nn.Conv2d(channels, channels // red, 1),
            nn.ReLU(inplace=True),
            nn.Conv2d(channels // red, channels, 1),
            nn.Sigmoid()
        )
    def forward(self, x):
        return x * self.fc(x)

class EfficientNetSE(nn.Module):
    def __init__(self, backbone, num_classes, drop=0.3):
        super().__init__()
        self.backbone = backbone
        self.se = SEBlock(1280)
        self.pool = nn.AdaptiveAvgPool2d(1)
        self.classifier = nn.Sequential(
            nn.Dropout(drop),
            nn.Linear(1280, num_classes)
        )
    def forward(self, x):
        x = self.backbone.features(x)
        x = self.se(x)
        x = self.pool(x).flatten(1)
        return self.classifier(x)

# ==== PREPROCESADO ====
def load_and_normalize(path, sr=SR, target_dBFS=-20.0):
    y, _ = librosa.load(path, sr=sr)
    y = y - np.mean(y)
    rms = np.sqrt(np.mean(y ** 2)) + 1e-9
    scalar = (10 ** (target_dBFS / 20)) / rms
    return y * scalar

def bandpass(y, sr=SR, low=FMIN, high=FMAX, order=6):
    from scipy.signal import butter, filtfilt
    nyq = 0.5 * sr
    b, a = butter(order, [low / nyq, high / nyq], btype='band')
    return filtfilt(b, a, y)

def segment(y, sr=SR, win=DURATION, hop=1.0):
    w = int(win * sr)
    h = int(hop * sr)
    if len(y) < w:
        y = np.pad(y, (0, w - len(y)))
        return [y]
    return [y[i:i + w] for i in range(0, len(y) - w + 1, h)]

def extract_log_mel(y, sr=SR, n_mels=BANDS, hop_length=HOP, fmin=FMIN, fmax=FMAX):
    mel = librosa.feature.melspectrogram(
        y=y, sr=sr, n_mels=n_mels, hop_length=hop_length, fmin=fmin, fmax=fmax, power=1.0)
    pcen = librosa.pcen(mel * (2 ** 31))
    return pcen

# ==== PREDICCIÓN SEGMENTADA ====
def predict_segments(file_path, model):
    y = load_and_normalize(file_path)
    y = bandpass(y, SR)
    segments = segment(y, SR)
    all_probs = []
    model.eval()
    with torch.no_grad():
        for seg in segments:
            mel = extract_log_mel(seg)
            inp = torch.tensor(mel[None, None], dtype=torch.float32).to(DEVICE)
            probs = torch.sigmoid(model(inp)).cpu().numpy()[0]
            all_probs.append(probs)
    return np.array(all_probs)

# ==== ESTRATEGIA HÍBRIDA DE PREDICCIÓN ====
def predict_file_with_hybrid_strategy(file_path, model, thresholds, label_encoder, override_max=0.9):
    probs = predict_segments(file_path, model)
    mean_probs = probs.mean(axis=0)
    max_probs = probs.max(axis=0)
    sensitive_thresh = [t - 0.15 for t in thresholds]

    preds = []
    for i, sp in enumerate(label_encoder.classes_):
        if mean_probs[i] > sensitive_thresh[i] or max_probs[i] > override_max:
            preds.append(sp)
    return preds, mean_probs, max_probs, probs

# ==== MAIN ====
if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("audio_file", type=str, help="Ruta al archivo de audio (.wav)")
    parser.add_argument("--model", default="CNN_final.pth", help="Ruta al modelo CNN .pth")
    parser.add_argument("--meta", default="label_encoder_and_thresholds.pkl", help="Pickle con encoder y thresholds")
    args = parser.parse_args()

    # Cargar metadatos (label encoder, thresholds, calibrators si los quieres aplicar también)
    with open(args.meta, "rb") as f:
        meta = pickle.load(f)

    label_encoder = meta["label_encoder"]
    thresholds = meta["thresholds"]

    # Cargar modelo
    from torchvision import models
    backbone = models.efficientnet_b0(weights=None)
    backbone.features[0][0] = nn.Conv2d(1, 32, kernel_size=3, stride=2, padding=1, bias=False)
    model = EfficientNetSE(backbone, num_classes=len(label_encoder.classes_))
    model.load_state_dict(torch.load(args.model, map_location=DEVICE))
    model.to(DEVICE)

    # Ejecutar predicción
    file_path = args.audio_file
    preds, mean_probs, max_probs, probs_all = predict_file_with_hybrid_strategy(
        file_path, model, thresholds, label_encoder
    )

    print(f"\n Archivo: {file_path}")
    print(f"Especies detectadas: {', '.join(preds)}\n")

    print("📊 Probabilidades por especie:")
    for i, sp in enumerate(label_encoder.classes_):
        print(f"  {sp:<25} → mean: {mean_probs[i]:.2f}, max: {max_probs[i]:.2f}")