Script para deteccion de gatos

.py

import os
import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
from tensorflow.keras.preprocessing import image
from tensorflow.keras.applications.mobilenet_v2 import MobileNetV2, preprocess_input, decode_predictions

# -----------------------------
# CONFIGURACIONES
# -----------------------------
DATASET_DIR = "./imagenes caso D"  # Carpeta de imágenes
TARGET_CLASSES = ["cat", "tiger", "panther"]  # Palabras clave de las clases
IMG_SIZE = (224, 224)

# -----------------------------
# CARGAR MODELO PREENTRENADO
# -----------------------------
model = MobileNetV2(weights="imagenet")

# -----------------------------
# VARIABLES PARA RECUENTO
# -----------------------------
counts = {cls: 0 for cls in TARGET_CLASSES}

# -----------------------------
# PROCESAR IMÁGENES
# -----------------------------
for img_name in os.listdir(DATASET_DIR):
    img_path = os.path.join(DATASET_DIR, img_name)

    try:
        # Cargar y preprocesar imagen
        img = image.load_img(img_path, target_size=IMG_SIZE)
        x = image.img_to_array(img)
        x = np.expand_dims(x, axis=0)
        x = preprocess_input(x)

        # Predicción
        preds = model.predict(x)
        decoded = decode_predictions(preds, top=3)[0]

        # Buscar si la predicción corresponde a nuestras clases objetivo
        found_class = None
        for _, label, prob in decoded:
            label = label.lower()
            if "cat" in label and "wildcat" not in label:  # excluir wildcat
                found_class = "cat"
                break
            elif "tiger" in label:
                found_class = "tiger"
                break
            elif "panther" in label or "jaguar" in label or "leopard" in label:
                found_class = "panther"
                break

        if found_class:
            counts[found_class] += 1
            print(f"{img_name} -> {found_class}")
        else:
            print(f"{img_name} -> No identificado en {TARGET_CLASSES}")

    except Exception as e:
        print(f"Error con {img_name}: {e}")

# -----------------------------
# MOSTRAR RECUENTO FINAL
# -----------------------------
print("\nRecuento final:")
for animal, count in counts.items():
    print(f"{animal}: {count}")

# -----------------------------
# GRAFICAR
# -----------------------------
plt.bar(counts.keys(), counts.values())
plt.title("Recuento de animales detectados")
plt.xlabel("Animal")
plt.ylabel("Cantidad")
plt.show()