File size: 2,998 Bytes
a2b861c d617c7d a2b861c d617c7d |
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 |
import pickle
import numpy as np
import io
import math
from fastapi import FastAPI, File, UploadFile, HTTPException
from PIL import Image
app = FastAPI(
title="Peruvian Sign Language (LSP) Recognition API",
description="Sube una imagen de una seña del alfabeto de la LSP para predecir la letra correspondiente usando un Mapa Autoorganizado (SOM).",
version="1.0.0"
)
try:
with open('lsp_som_model.pkl', 'rb') as f:
model_data = pickle.load(f)
som = model_data['som']
label_map = model_data['label_map']
CLASSES = model_data['classes'] # La lista ['A', 'B', 'C', ...]
IMG_SIZE = model_data['img_size'] # El tamaño de la imagen
print("✅ Modelo y activos cargados exitosamente.")
print(f" - Clases reconocidas: {CLASSES}")
print(f" - Tamaño de imagen esperado: {IMG_SIZE}x{IMG_SIZE}")
except FileNotFoundError:
print("❌ ERROR: No se encontró el archivo del modelo 'lsp_som_model.pkl'.")
som = None
def preprocess_image_from_bytes(image_bytes: bytes):
try:
img = Image.open(io.BytesIO(image_bytes)).convert('L') # 'L' para escala de grises
img = img.resize((IMG_SIZE, IMG_SIZE))
img_array = np.array(img)
img_normalized = img_array / 255.0
return img_normalized.flatten()
except Exception as e:
raise HTTPException(status_code=400, detail=f"Archivo de imagen inválido. Error: {e}")
@app.get("/", tags=["Status"])
def read_root():
return {"status": "ok", "message": "API de Reconocimiento de LSP!!"}
@app.post("/predict", tags=["Prediction"])
async def predict_sign(file: UploadFile = File(..., description="Un archivo de imagen de una seña de la LSP.")):
if not som:
raise HTTPException(status_code=503, detail="El modelo no está cargado.")
image_bytes = await file.read()
feature_vector = preprocess_image_from_bytes(image_bytes)
winner_neuron = som.winner(feature_vector)
predicted_index = label_map.get(winner_neuron, -1)
# Vecino mas cercano para prediccion
is_best_guess = False
if predicted_index == -1:
is_best_guess = True
min_dist = float('inf')
for mapped_pos, mapped_label in label_map.items():
dist = math.sqrt((winner_neuron[0] - mapped_pos[0])**2 + (winner_neuron[1] - mapped_pos[1])**2)
if dist < min_dist:
min_dist = dist
predicted_index = mapped_label
if predicted_index != -1:
predicted_letter = CLASSES[predicted_index]
prediction_type = "Nearest Neighbor" if is_best_guess else "Direct Match"
else:
predicted_letter = "Unknown"
prediction_type = "Error (No Mapped Neurons Found)"
response = {
"filename": file.filename,
"predicted_letter": predicted_letter,
"prediction_type": prediction_type,
"winner_neuron_on_map": [int(coord) for coord in winner_neuron]
}
print(f"[LOG] Respuesta enviada: {response}")
return response |