from PIL import Image from PIL import ImageEnhance from PIL import ImageFilter import cv2 import numpy as np import sys # Arguments input_path = sys.argv[1] output_path = sys.argv[2] scale = sys.argv[3] if len(sys.argv) > 3 else "2x" # Load image img = Image.open(input_path) # Convert RGB if img.mode != "RGB": img = img.convert("RGB") # Original size w, h = img.size print("Scale:", scale) print("Original:", w, h) # ========================================================== # 2X MODE # ========================================================== if scale == "2x": img = img.resize( (w * 2, h * 2), Image.LANCZOS ) img = img.filter( ImageFilter.SHARPEN ) contrast = ImageEnhance.Contrast(img) img = contrast.enhance(1.2) color = ImageEnhance.Color(img) img = color.enhance(1.1) brightness = ImageEnhance.Brightness(img) img = brightness.enhance(1.05) sharpness = ImageEnhance.Sharpness(img) img = sharpness.enhance(1.3) img.save( output_path, quality=95, optimize=True ) # ============ # NOISE REMOVAL # ============ elif scale == "noise": img = cv2.imread(input_path) if img is None: print(f"[ERROR] Failed to load image: {input_path}") sys.exit(1) # ── Step 1: Pre-smoothing — larger kernel for full-body grain suppression img = cv2.GaussianBlur(img, (5, 5), sigmaX=1.2) # ── Step 2: Strong full-image denoising img = cv2.fastNlMeansDenoisingColored( img, None, h=18, hColor=15, templateWindowSize=9, searchWindowSize=27, ) # ── Step 3: Bilateral filter — edge-aware pass over entire image img = cv2.bilateralFilter( img, d=0, sigmaColor=0, sigmaSpace=65, ) # ── Step 4: CLAHE — gentle contrast recovery across full image lab = cv2.cvtColor(img, cv2.COLOR_BGR2LAB) l_channel, a_channel, b_channel = cv2.split(lab) clahe = cv2.createCLAHE( clipLimit=1.2, tileGridSize=(8, 8), ) l_enhanced = clahe.apply(l_channel) img = cv2.cvtColor(cv2.merge((l_enhanced, a_channel, b_channel)), cv2.COLOR_LAB2BGR) # ── Step 5: Lighter sharpening — recover edges without re-introducing grain SHARPEN_KERNEL = np.array([ [ 0, -0.8, 0], [-0.8, 4.2, -0.8], # ⬇ was 4.8 — softer to avoid re-sharpening noise in fabric [ 0, -0.8, 0], ], dtype=np.float32) img = cv2.filter2D(img, ddepth=-1, kernel=SHARPEN_KERNEL) # ── Write result success = cv2.imwrite(output_path, img) if not success: print(f"[ERROR] Failed to write output: {output_path}") sys.exit(1) # ========================================================== # 4X MODE # ========================================================== elif scale == "4x": multiplier = 2.5 new_w = int(w * multiplier) new_h = int(h * multiplier) img = img.resize( (new_w, new_h), Image.LANCZOS ) img = img.filter( ImageFilter.SHARPEN ) sharpness = ImageEnhance.Sharpness(img) img = sharpness.enhance(2.0) contrast = ImageEnhance.Contrast(img) img = contrast.enhance(1.4) color = ImageEnhance.Color(img) img = color.enhance(1.2) brightness = ImageEnhance.Brightness(img) img = brightness.enhance(1.08) img = img.filter( ImageFilter.DETAIL ) img.save( output_path, quality=95, optimize=True ) # ========================================================== # FACE ENHANCE # ========================================================== elif scale == "face": img_cv = cv2.imread(input_path) if img_cv is None: print("Image load failed") sys.exit(1) # upscale slightly img_cv = cv2.resize( img_cv, None, fx=1.8, fy=1.8, interpolation=cv2.INTER_CUBIC ) # whole image denoise img_cv = cv2.fastNlMeansDenoisingColored( img_cv, None, 4, 4, 7, 21 ) # subtle sharpening kernel = np.array([ [0, -1, 0], [-1, 5.2, -1], [0, -1, 0] ]) img_cv = cv2.filter2D( img_cv, -1, kernel ) # improve contrast naturally lab = cv2.cvtColor( img_cv, cv2.COLOR_BGR2LAB ) l, a, b = cv2.split(lab) clahe = cv2.createCLAHE( clipLimit=1.8, tileGridSize=(8,8) ) cl = clahe.apply(l) merged = cv2.merge((cl, a, b)) img_cv = cv2.cvtColor( merged, cv2.COLOR_LAB2BGR ) # smooth very slightly img_cv = cv2.bilateralFilter( img_cv, 5, 30, 30 ) cv2.imwrite( output_path, img_cv ) # ============= # PORTRAIT # ============= elif scale == "portrait": img_cv = cv2.imread(input_path) img_cv = cv2.resize( img_cv, None, fx=1.5, fy=1.5, interpolation=cv2.INTER_CUBIC ) # smooth slightly img_cv = cv2.bilateralFilter( img_cv, 7, 50, 50 ) # sharpen lightly kernel = np.array([ [0,-1,0], [-1,5,-1], [0,-1,0] ]) img_cv = cv2.filter2D( img_cv, -1, kernel ) # improve colors hsv = cv2.cvtColor( img_cv, cv2.COLOR_BGR2HSV ) hsv[:,:,1] = cv2.multiply( hsv[:,:,1], 1.15 ) img_cv = cv2.cvtColor( hsv, cv2.COLOR_HSV2BGR ) cv2.imwrite(output_path, img_cv) # =========== # BLUR FIX # =========== elif scale == "blurfix": img_cv = cv2.imread(input_path) img_cv = cv2.resize( img_cv, None, fx=2, fy=2, interpolation=cv2.INTER_CUBIC ) # strong sharpen kernel = np.array([ [0,-1,0], [-1,6,-1], [0,-1,0] ]) img_cv = cv2.filter2D( img_cv, -1, kernel ) # denoise img_cv = cv2.fastNlMeansDenoisingColored( img_cv, None, 5, 5, 7, 21 ) cv2.imwrite(output_path, img_cv) # ========== # CARTOON # ========== elif scale == "cartoon": img_cv = cv2.imread(input_path) gray = cv2.cvtColor( img_cv, cv2.COLOR_BGR2GRAY ) gray = cv2.medianBlur(gray, 5) edges = cv2.adaptiveThreshold( gray, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 9, 9 ) color = cv2.bilateralFilter( img_cv, 9, 250, 250 ) cartoon = cv2.bitwise_and( color, color, mask=edges ) cv2.imwrite(output_path, cartoon) # ============ # ANIME # ============ elif scale == "anime": img_cv = cv2.imread(input_path) # smooth colors for i in range(2): img_cv = cv2.bilateralFilter( img_cv, 9, 75, 75 ) # edges gray = cv2.cvtColor( img_cv, cv2.COLOR_BGR2GRAY ) blur = cv2.medianBlur(gray, 5) edges = cv2.adaptiveThreshold( blur, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 9, 2 ) anime = cv2.bitwise_and( img_cv, img_cv, mask=edges ) # boost saturation hsv = cv2.cvtColor( anime, cv2.COLOR_BGR2HSV ) hsv[:,:,1] = cv2.multiply( hsv[:,:,1], 1.4 ) anime = cv2.cvtColor( hsv, cv2.COLOR_HSV2BGR ) cv2.imwrite(output_path, anime) # ============= # HD # ============= elif scale == "hd": img = img.resize( (w * 2, h * 2), Image.LANCZOS ) # sharpness sharpness = ImageEnhance.Sharpness(img) img = sharpness.enhance(2.2) # contrast contrast = ImageEnhance.Contrast(img) img = contrast.enhance(1.35) # color color = ImageEnhance.Color(img) img = color.enhance(1.25) # detail img = img.filter( ImageFilter.DETAIL ) img.save( output_path, quality=98, optimize=True ) # ========================================================== # DEFAULT # ========================================================== else: img.save(output_path) print("DONE")