GPU Accelerated OpenCV

modules/capturer.py

@@ -1,6 +1,7 @@
 from typing import Any
 import cv2
 import modules.globals  # Import the globals to check the color correction toggle
+from modules.gpu_processing import gpu_cvt_color
 
 
 def get_video_frame(video_path: str, frame_number: int = 0) -> Any:
@@ -19,7 +20,7 @@ def get_video_frame(video_path: str, frame_number: int = 0) -> Any:
 
     if has_frame and modules.globals.color_correction:
         # Convert the frame color if necessary
-        frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
+        frame = gpu_cvt_color(frame, cv2.COLOR_BGR2RGB)
 
     capture.release()
     return frame if has_frame else None

modules/gpu_processing.py

@@ -0,0 +1,286 @@
+# --- START OF FILE gpu_processing.py ---
+"""
+GPU-accelerated image processing using OpenCV CUDA (cv2.cuda.GpuMat).
+
+Provides drop-in replacements for common cv2 functions.  When OpenCV is built
+with CUDA support the functions transparently upload → process → download via
+GpuMat; otherwise they fall back to the regular CPU path so the rest of the
+codebase never has to care whether CUDA is available.
+
+Usage
+-----
+    from modules.gpu_processing import (
+        gpu_gaussian_blur, gpu_sharpen, gpu_add_weighted,
+        gpu_resize, gpu_cvt_color, gpu_flip,
+        is_gpu_accelerated,
+    )
+"""
+
+from __future__ import annotations
+
+import cv2
+import numpy as np
+from typing import Tuple, Optional
+
+# ---------------------------------------------------------------------------
+# CUDA availability detection (evaluated once at import time)
+# ---------------------------------------------------------------------------
+CUDA_AVAILABLE: bool = False
+
+try:
+    # cv2.cuda.GpuMat is only present when OpenCV is compiled with CUDA
+    _test_mat = cv2.cuda.GpuMat()
+    # Verify we have the required filter / image-processing functions
+    _has_gauss = hasattr(cv2.cuda, "createGaussianFilter")
+    _has_resize = hasattr(cv2.cuda, "resize")
+    _has_cvt = hasattr(cv2.cuda, "cvtColor")
+    if _has_gauss and _has_resize and _has_cvt:
+        CUDA_AVAILABLE = True
+        print("[gpu_processing] OpenCV CUDA support detected – GPU-accelerated processing enabled.")
+    else:
+        missing = []
+        if not _has_gauss:
+            missing.append("createGaussianFilter")
+        if not _has_resize:
+            missing.append("resize")
+        if not _has_cvt:
+            missing.append("cvtColor")
+        print(f"[gpu_processing] cv2.cuda.GpuMat exists but missing: {', '.join(missing)} – falling back to CPU.")
+except Exception:
+    print("[gpu_processing] OpenCV CUDA not available – using CPU fallback for all operations.")
+
+
+# ---------------------------------------------------------------------------
+# Internal helpers
+# ---------------------------------------------------------------------------
+
+def _ensure_uint8(img: np.ndarray) -> np.ndarray:
+    """Clip and convert to uint8 if necessary."""
+    if img.dtype != np.uint8:
+        return np.clip(img, 0, 255).astype(np.uint8)
+    return img
+
+
+def _ksize_odd(ksize: Tuple[int, int]) -> Tuple[int, int]:
+    """Ensure kernel dimensions are positive and odd (required by GaussianBlur)."""
+    kw = max(1, ksize[0] // 2 * 2 + 1) if ksize[0] > 0 else 0
+    kh = max(1, ksize[1] // 2 * 2 + 1) if ksize[1] > 0 else 0
+    return (kw, kh)
+
+
+def _cv_type_for(img: np.ndarray) -> int:
+    """Return the OpenCV type constant matching *img* (uint8 only)."""
+    channels = 1 if img.ndim == 2 else img.shape[2]
+    if channels == 1:
+        return cv2.CV_8UC1
+    elif channels == 3:
+        return cv2.CV_8UC3
+    elif channels == 4:
+        return cv2.CV_8UC4
+    return cv2.CV_8UC3  # fallback
+
+
+# ---------------------------------------------------------------------------
+# Public API – Gaussian Blur
+# ---------------------------------------------------------------------------
+
+def gpu_gaussian_blur(
+    src: np.ndarray,
+    ksize: Tuple[int, int],
+    sigma_x: float,
+    sigma_y: float = 0,
+) -> np.ndarray:
+    """Drop-in replacement for ``cv2.GaussianBlur`` with CUDA acceleration.
+
+    Parameters match ``cv2.GaussianBlur(src, ksize, sigmaX, sigmaY)``.
+    When *ksize* is ``(0, 0)`` OpenCV computes the kernel size from *sigma_x*.
+    """
+    if CUDA_AVAILABLE:
+        try:
+            src_u8 = _ensure_uint8(src)
+            cv_type = _cv_type_for(src_u8)
+            ks = _ksize_odd(ksize) if ksize != (0, 0) else ksize
+
+            gauss = cv2.cuda.createGaussianFilter(cv_type, cv_type, ks, sigma_x, sigma_y)
+            gpu_src = cv2.cuda.GpuMat()
+            gpu_src.upload(src_u8)
+            gpu_dst = gauss.apply(gpu_src)
+            return gpu_dst.download()
+        except cv2.error:
+            pass
+
+    return cv2.GaussianBlur(src, ksize, sigma_x, sigmaY=sigma_y)
+
+
+# ---------------------------------------------------------------------------
+# Public API – addWeighted
+# ---------------------------------------------------------------------------
+
+def gpu_add_weighted(
+    src1: np.ndarray,
+    alpha: float,
+    src2: np.ndarray,
+    beta: float,
+    gamma: float,
+) -> np.ndarray:
+    """Drop-in replacement for ``cv2.addWeighted`` with CUDA acceleration."""
+    if CUDA_AVAILABLE:
+        try:
+            s1 = _ensure_uint8(src1)
+            s2 = _ensure_uint8(src2)
+            g1 = cv2.cuda.GpuMat()
+            g2 = cv2.cuda.GpuMat()
+            g1.upload(s1)
+            g2.upload(s2)
+            gpu_dst = cv2.cuda.addWeighted(g1, alpha, g2, beta, gamma)
+            return gpu_dst.download()
+        except cv2.error:
+            pass
+
+    return cv2.addWeighted(src1, alpha, src2, beta, gamma)
+
+
+# ---------------------------------------------------------------------------
+# Public API – Unsharp-mask sharpening
+# ---------------------------------------------------------------------------
+
+def gpu_sharpen(
+    src: np.ndarray,
+    strength: float,
+    sigma: float = 3,
+) -> np.ndarray:
+    """Unsharp-mask sharpening, optionally GPU-accelerated.
+
+    Equivalent to::
+
+        blurred = GaussianBlur(src, (0,0), sigma)
+        result  = addWeighted(src, 1+strength, blurred, -strength, 0)
+    """
+    if strength <= 0:
+        return src
+
+    if CUDA_AVAILABLE:
+        try:
+            src_u8 = _ensure_uint8(src)
+            cv_type = _cv_type_for(src_u8)
+
+            gauss = cv2.cuda.createGaussianFilter(cv_type, cv_type, (0, 0), sigma)
+            gpu_src = cv2.cuda.GpuMat()
+            gpu_src.upload(src_u8)
+            gpu_blurred = gauss.apply(gpu_src)
+            gpu_sharp = cv2.cuda.addWeighted(gpu_src, 1.0 + strength, gpu_blurred, -strength, 0)
+            result = gpu_sharp.download()
+            return np.clip(result, 0, 255).astype(np.uint8)
+        except cv2.error:
+            pass
+
+    blurred = cv2.GaussianBlur(src, (0, 0), sigma)
+    sharpened = cv2.addWeighted(src, 1.0 + strength, blurred, -strength, 0)
+    return np.clip(sharpened, 0, 255).astype(np.uint8)
+
+
+# ---------------------------------------------------------------------------
+# Public API – Resize
+# ---------------------------------------------------------------------------
+
+# Map common cv2 interpolation flags to their CUDA equivalents
+_INTERP_MAP = {
+    cv2.INTER_NEAREST: cv2.INTER_NEAREST,
+    cv2.INTER_LINEAR: cv2.INTER_LINEAR,
+    cv2.INTER_CUBIC: cv2.INTER_CUBIC,
+    cv2.INTER_AREA: cv2.INTER_AREA,
+    cv2.INTER_LANCZOS4: cv2.INTER_LANCZOS4,
+}
+
+
+def gpu_resize(
+    src: np.ndarray,
+    dsize: Tuple[int, int],
+    fx: float = 0,
+    fy: float = 0,
+    interpolation: int = cv2.INTER_LINEAR,
+) -> np.ndarray:
+    """Drop-in replacement for ``cv2.resize`` with CUDA acceleration.
+
+    Parameters match ``cv2.resize(src, dsize, fx=fx, fy=fy, interpolation=...)``.
+    """
+    if CUDA_AVAILABLE:
+        try:
+            src_u8 = _ensure_uint8(src)
+            gpu_src = cv2.cuda.GpuMat()
+            gpu_src.upload(src_u8)
+
+            interp = _INTERP_MAP.get(interpolation, cv2.INTER_LINEAR)
+
+            if dsize and dsize[0] > 0 and dsize[1] > 0:
+                gpu_dst = cv2.cuda.resize(gpu_src, dsize, interpolation=interp)
+            else:
+                gpu_dst = cv2.cuda.resize(gpu_src, (0, 0), fx=fx, fy=fy, interpolation=interp)
+
+            return gpu_dst.download()
+        except cv2.error:
+            pass
+
+    return cv2.resize(src, dsize, fx=fx, fy=fy, interpolation=interpolation)
+
+
+# ---------------------------------------------------------------------------
+# Public API – Color conversion
+# ---------------------------------------------------------------------------
+
+def gpu_cvt_color(
+    src: np.ndarray,
+    code: int,
+) -> np.ndarray:
+    """Drop-in replacement for ``cv2.cvtColor`` with CUDA acceleration.
+
+    Parameters match ``cv2.cvtColor(src, code)``.
+    """
+    if CUDA_AVAILABLE:
+        try:
+            src_u8 = _ensure_uint8(src)
+            gpu_src = cv2.cuda.GpuMat()
+            gpu_src.upload(src_u8)
+            gpu_dst = cv2.cuda.cvtColor(gpu_src, code)
+            return gpu_dst.download()
+        except cv2.error:
+            pass
+
+    return cv2.cvtColor(src, code)
+
+
+# ---------------------------------------------------------------------------
+# Public API – Flip
+# ---------------------------------------------------------------------------
+
+def gpu_flip(
+    src: np.ndarray,
+    flip_code: int,
+) -> np.ndarray:
+    """Drop-in replacement for ``cv2.flip`` with CUDA acceleration.
+
+    Parameters match ``cv2.flip(src, flipCode)``.
+    *flip_code*: 0 = vertical, 1 = horizontal, -1 = both.
+    """
+    if CUDA_AVAILABLE:
+        try:
+            src_u8 = _ensure_uint8(src)
+            gpu_src = cv2.cuda.GpuMat()
+            gpu_src.upload(src_u8)
+            gpu_dst = cv2.cuda.flip(gpu_src, flip_code)
+            return gpu_dst.download()
+        except cv2.error:
+            pass
+
+    return cv2.flip(src, flip_code)
+
+
+# ---------------------------------------------------------------------------
+# Convenience: check at runtime whether GPU path is active
+# ---------------------------------------------------------------------------
+
+def is_gpu_accelerated() -> bool:
+    """Return ``True`` when the CUDA path will be used."""
+    return CUDA_AVAILABLE
+
+# --- END OF FILE gpu_processing.py ---

modules/predicter.py

@@ -3,6 +3,7 @@ import opennsfw2
 from PIL import Image
 import cv2  # Add OpenCV import
 import modules.globals  # Import globals to access the color correction toggle
+from modules.gpu_processing import gpu_cvt_color
 
 from modules.typing import Frame
 
@@ -14,7 +15,7 @@ model = None
 def predict_frame(target_frame: Frame) -> bool:
     # Convert the frame to RGB before processing if color correction is enabled
     if modules.globals.color_correction:
-        target_frame = cv2.cvtColor(target_frame, cv2.COLOR_BGR2RGB)
+        target_frame = gpu_cvt_color(target_frame, cv2.COLOR_BGR2RGB)
         
     image = Image.fromarray(target_frame)
     image = opennsfw2.preprocess_image(image, opennsfw2.Preprocessing.YAHOO)

modules/processors/frame/face_masking.py

@@ -2,6 +2,7 @@ import cv2
 import numpy as np
 from modules.typing import Face, Frame
 import modules.globals
+from modules.gpu_processing import gpu_gaussian_blur, gpu_resize, gpu_cvt_color
 
 def apply_color_transfer(source, target):
     """
@@ -61,8 +62,8 @@ def create_face_mask(face: Face, frame: Frame) -> np.ndarray:
         # Fill the padded convex hull
         cv2.fillConvexPoly(mask, hull_padded, 255)
 
-        # Smooth the mask edges
-        mask = cv2.GaussianBlur(mask, (5, 5), 3)
+        # Smooth the mask edges (GPU-accelerated when available)
+        mask = gpu_gaussian_blur(mask, (5, 5), 3)
 
     return mask
 
@@ -123,8 +124,8 @@ def create_lower_mouth_mask(
         polygon_relative_to_roi = expanded_landmarks - [min_x, min_y]
         cv2.fillPoly(mask_roi, [polygon_relative_to_roi], 255)
 
-        # Apply Gaussian blur to soften the mask edges
-        mask_roi = cv2.GaussianBlur(mask_roi, (15, 15), 5)
+        # Apply Gaussian blur to soften the mask edges (GPU-accelerated when available)
+        mask_roi = gpu_gaussian_blur(mask_roi, (15, 15), 5)
 
         # Place the mask ROI in the full-sized mask
         mask[min_y:max_y, min_x:max_x] = mask_roi
@@ -192,8 +193,8 @@ def create_eyes_mask(face: Face, frame: Frame) -> (np.ndarray, np.ndarray, tuple
         cv2.ellipse(mask_roi, left_center, left_axes, 0, 0, 360, 255, -1)
         cv2.ellipse(mask_roi, right_center, right_axes, 0, 0, 360, 255, -1)
         
-        # Apply Gaussian blur to soften mask edges
-        mask_roi = cv2.GaussianBlur(mask_roi, (15, 15), 5)
+        # Apply Gaussian blur to soften mask edges (GPU-accelerated when available)
+        mask_roi = gpu_gaussian_blur(mask_roi, (15, 15), 5)
         
         # Place the mask ROI in the full-sized mask
         mask[min_y:max_y, min_x:max_x] = mask_roi
@@ -374,15 +375,15 @@ def create_eyebrows_mask(face: Face, frame: Frame) -> (np.ndarray, np.ndarray, t
             left_shape = create_curved_eyebrow(left_local)
             right_shape = create_curved_eyebrow(right_local)
             
-            # Apply multi-stage blurring for natural feathering
+            # Apply multi-stage blurring for natural feathering (GPU-accelerated when available)
             # First, strong Gaussian blur for initial softening
-            mask_roi = cv2.GaussianBlur(mask_roi, (21, 21), 7)
+            mask_roi = gpu_gaussian_blur(mask_roi, (21, 21), 7)
             
             # Second, medium blur for transition areas
-            mask_roi = cv2.GaussianBlur(mask_roi, (11, 11), 3)
+            mask_roi = gpu_gaussian_blur(mask_roi, (11, 11), 3)
             
             # Finally, light blur for fine details
-            mask_roi = cv2.GaussianBlur(mask_roi, (5, 5), 1)
+            mask_roi = gpu_gaussian_blur(mask_roi, (5, 5), 1)
             
             # Normalize mask values
             mask_roi = cv2.normalize(mask_roi, None, 0, 255, cv2.NORM_MINMAX)
@@ -405,7 +406,7 @@ def create_eyebrows_mask(face: Face, frame: Frame) -> (np.ndarray, np.ndarray, t
             right_local = right_eyebrow - [min_x, min_y]
             cv2.fillPoly(mask_roi, [left_local.astype(np.int32)], 255)
             cv2.fillPoly(mask_roi, [right_local.astype(np.int32)], 255)
-            mask_roi = cv2.GaussianBlur(mask_roi, (21, 21), 7)
+            mask_roi = gpu_gaussian_blur(mask_roi, (21, 21), 7)
             mask[min_y:max_y, min_x:max_x] = mask_roi
             eyebrows_cutout = frame[min_y:max_y, min_x:max_x].copy()
             eyebrows_polygon = np.vstack([left_eyebrow, right_eyebrow]).astype(np.int32)
@@ -433,11 +434,11 @@ def apply_mask_area(
         return frame
 
     try:
-        resized_cutout = cv2.resize(cutout, (box_width, box_height))
+        resized_cutout = gpu_resize(cutout, (box_width, box_height))
         roi = frame[min_y:max_y, min_x:max_x]
 
         if roi.shape != resized_cutout.shape:
-            resized_cutout = cv2.resize(
+            resized_cutout = gpu_resize(
                 resized_cutout, (roi.shape[1], roi.shape[0])
             )
 
@@ -457,8 +458,8 @@ def apply_mask_area(
             adjusted_polygon = polygon - [min_x, min_y]
             cv2.fillPoly(polygon_mask, [adjusted_polygon], 255)
 
-        # Apply strong initial feathering
-        polygon_mask = cv2.GaussianBlur(polygon_mask, (21, 21), 7)
+        # Apply strong initial feathering (GPU-accelerated when available)
+        polygon_mask = gpu_gaussian_blur(polygon_mask, (21, 21), 7)
 
         # Apply additional feathering
         feather_amount = min(

modules/processors/frame/face_swapper.py

@@ -15,6 +15,7 @@ from modules.utilities import (
     is_video,
 )
 from modules.cluster_analysis import find_closest_centroid
+from modules.gpu_processing import gpu_gaussian_blur, gpu_sharpen, gpu_add_weighted, gpu_resize, gpu_cvt_color
 import os
 from collections import deque
 import time
@@ -158,7 +159,7 @@ def swap_face(source_face: Face, target_face: Face, temp_frame: Frame) -> Frame:
              # print(f"Warning: Swapped frame shape {swapped_frame_raw.shape} differs from input {temp_frame.shape}.") # Debug
              # Attempt resize (might distort if aspect ratio changed, but better than crashing)
              try:
-                 swapped_frame_raw = cv2.resize(swapped_frame_raw, (temp_frame.shape[1], temp_frame.shape[0]))
+                 swapped_frame_raw = gpu_resize(swapped_frame_raw, (temp_frame.shape[1], temp_frame.shape[0]))
              except Exception as resize_e:
                  # print(f"Error resizing swapped frame: {resize_e}") # Debug
                  return original_frame
@@ -236,7 +237,7 @@ def swap_face(source_face: Face, target_face: Face, temp_frame: Frame) -> Frame:
 
     # Blend the original_frame with the (potentially mouth-masked) swapped_frame
     # Ensure both frames are uint8 before blending
-    final_swapped_frame = cv2.addWeighted(original_frame.astype(np.uint8), 1 - opacity, swapped_frame.astype(np.uint8), opacity, 0)
+    final_swapped_frame = gpu_add_weighted(original_frame.astype(np.uint8), 1 - opacity, swapped_frame.astype(np.uint8), opacity, 0)
 
     # Ensure final frame is uint8 after blending (addWeighted should preserve it, but belt-and-suspenders)
     final_swapped_frame = final_swapped_frame.astype(np.uint8)
@@ -312,17 +313,10 @@ def apply_post_processing(current_frame: Frame, swapped_face_bboxes: List[np.nda
             face_region = processed_frame[y1:y2, x1:x2]
             if face_region.size == 0: continue
 
-            # Apply sharpening with optimized parameters for Apple Silicon
+            # Apply sharpening (GPU-accelerated when CUDA OpenCV is available)
             try:
-                # Use smaller sigma for faster processing on Apple Silicon
                 sigma = 2 if IS_APPLE_SILICON else 3
-                blurred = cv2.GaussianBlur(face_region, (0, 0), sigma)
-                sharpened_region = cv2.addWeighted(
-                    face_region, 1.0 + sharpness_value,
-                    blurred, -sharpness_value,
-                    0
-                )
-                sharpened_region = np.clip(sharpened_region, 0, 255).astype(np.uint8)
+                sharpened_region = gpu_sharpen(face_region, strength=sharpness_value, sigma=sigma)
                 processed_frame[y1:y2, x1:x2] = sharpened_region
             except cv2.error:
                 pass
@@ -338,7 +332,7 @@ def apply_post_processing(current_frame: Frame, swapped_face_bboxes: List[np.nda
         if PREVIOUS_FRAME_RESULT is not None and PREVIOUS_FRAME_RESULT.shape == processed_frame.shape and PREVIOUS_FRAME_RESULT.dtype == processed_frame.dtype:
             # Perform interpolation
             try:
-                 final_frame = cv2.addWeighted(
+                 final_frame = gpu_add_weighted(
                     PREVIOUS_FRAME_RESULT, 1.0 - interpolation_weight,
                     processed_frame, interpolation_weight,
                     0
@@ -813,10 +807,10 @@ def create_lower_mouth_mask(
             # Draw polygon on the ROI mask
             cv2.fillPoly(mask_roi, [polygon_relative_to_roi], 255)
 
-            # Apply Gaussian blur (ensure kernel size is odd and positive)
+            # Apply Gaussian blur (GPU-accelerated when available)
             blur_k_size = getattr(modules.globals, "mask_blur_kernel", 15) # Default 15
             blur_k_size = max(1, blur_k_size // 2 * 2 + 1) # Ensure odd
-            mask_roi = cv2.GaussianBlur(mask_roi, (blur_k_size, blur_k_size), 0) # Sigma=0 calculates from kernel
+            mask_roi = gpu_gaussian_blur(mask_roi, (blur_k_size, blur_k_size), 0)
 
             # Place the mask ROI in the full-sized mask
             mask[min_y:max_y, min_x:max_x] = mask_roi
@@ -952,7 +946,7 @@ def apply_mouth_area(
         if roi.shape[:2] != mouth_cutout.shape[:2]:
              # Check if mouth_cutout has valid dimensions before resizing
              if mouth_cutout.shape[0] > 0 and mouth_cutout.shape[1] > 0:
-                 resized_mouth_cutout = cv2.resize(mouth_cutout, (box_width, box_height), interpolation=cv2.INTER_LINEAR)
+                  resized_mouth_cutout = gpu_resize(mouth_cutout, (box_width, box_height), interpolation=cv2.INTER_LINEAR)
              else:
                  # print("Warning: mouth_cutout has invalid dimensions, cannot resize.")
                  return frame # Cannot proceed without valid cutout
@@ -1125,14 +1119,10 @@ def create_face_mask(face: Face, frame: Frame) -> np.ndarray:
              return mask # Return empty mask on error
 
 
-        # Apply Gaussian blur to feather the mask edges
-        # Kernel size should be reasonably large, odd, and positive
+        # Apply Gaussian blur to feather the mask edges (GPU-accelerated when available)
         blur_k_size = getattr(modules.globals, "face_mask_blur", 31) # Default 31
         blur_k_size = max(1, blur_k_size // 2 * 2 + 1) # Ensure odd and positive
-
-        # Use sigma=0 to let OpenCV calculate from kernel size
-        # Apply blur to the uint8 mask directly
-        mask = cv2.GaussianBlur(mask, (blur_k_size, blur_k_size), 0)
+        mask = gpu_gaussian_blur(mask, (blur_k_size, blur_k_size), 0)
 
         # --- Optional: Return float mask for apply_mouth_area ---
         # mask = mask.astype(float) / 255.0

modules/ui.py

@@ -4,6 +4,7 @@ import customtkinter as ctk
 from typing import Callable, Tuple
 import cv2
 from cv2_enumerate_cameras import enumerate_cameras  # Add this import
+from modules.gpu_processing import gpu_cvt_color, gpu_resize, gpu_flip
 from PIL import Image, ImageOps
 import time
 import json
@@ -546,7 +547,7 @@ def create_source_target_popup(
         )
         x_label.grid(row=id, column=2, padx=10, pady=10)
 
-        image = Image.fromarray(cv2.cvtColor(item["target"]["cv2"], cv2.COLOR_BGR2RGB))
+        image = Image.fromarray(gpu_cvt_color(item["target"]["cv2"], cv2.COLOR_BGR2RGB))
         image = image.resize(
             (MAPPER_PREVIEW_MAX_WIDTH, MAPPER_PREVIEW_MAX_HEIGHT), Image.LANCZOS
         )
@@ -601,7 +602,7 @@ def update_popup_source(
             }
 
             image = Image.fromarray(
-                cv2.cvtColor(map[button_num]["source"]["cv2"], cv2.COLOR_BGR2RGB)
+                gpu_cvt_color(map[button_num]["source"]["cv2"], cv2.COLOR_BGR2RGB)
             )
             image = image.resize(
                 (MAPPER_PREVIEW_MAX_WIDTH, MAPPER_PREVIEW_MAX_HEIGHT), Image.LANCZOS
@@ -794,7 +795,7 @@ def fit_image_to_size(image, width: int, height: int):
         ratio_w = width / w
     ratio = max(ratio_w, ratio_h)
     new_size = (int(ratio * w), int(ratio * h))
-    return cv2.resize(image, dsize=new_size)
+    return gpu_resize(image, dsize=new_size)
 
 
 def render_image_preview(image_path: str, size: Tuple[int, int]) -> ctk.CTkImage:
@@ -812,7 +813,7 @@ def render_video_preview(
         capture.set(cv2.CAP_PROP_POS_FRAMES, frame_number)
     has_frame, frame = capture.read()
     if has_frame:
-        image = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
+        image = Image.fromarray(gpu_cvt_color(frame, cv2.COLOR_BGR2RGB))
         if size:
             image = ImageOps.fit(image, size, Image.LANCZOS)
         return ctk.CTkImage(image, size=image.size)
@@ -850,7 +851,7 @@ def update_preview(frame_number: int = 0) -> None:
             temp_frame = frame_processor.process_frame(
                 get_one_face(cv2.imread(modules.globals.source_path)), temp_frame
             )
-        image = Image.fromarray(cv2.cvtColor(temp_frame, cv2.COLOR_BGR2RGB))
+        image = Image.fromarray(gpu_cvt_color(temp_frame, cv2.COLOR_BGR2RGB))
         image = ImageOps.contain(
             image, (PREVIEW_MAX_WIDTH, PREVIEW_MAX_HEIGHT), Image.LANCZOS
         )
@@ -1007,7 +1008,7 @@ def _processing_thread_func(capture_queue, processed_queue, stop_event):
         proc_frame_index += 1
 
         if modules.globals.live_mirror:
-            temp_frame = cv2.flip(temp_frame, 1)
+            temp_frame = gpu_flip(temp_frame, 1)
 
         if not modules.globals.map_faces:
             if source_image is None and modules.globals.source_path:
@@ -1136,7 +1137,7 @@ def create_webcam_preview(camera_index: int):
                 temp_frame, PREVIEW.winfo_width(), PREVIEW.winfo_height()
             )
 
-        image = cv2.cvtColor(temp_frame, cv2.COLOR_BGR2RGB)
+        image = gpu_cvt_color(temp_frame, cv2.COLOR_BGR2RGB)
         image = Image.fromarray(image)
         image = ImageOps.contain(
             image, (temp_frame.shape[1], temp_frame.shape[0]), Image.LANCZOS
@@ -1263,7 +1264,7 @@ def refresh_data(map: list):
 
         if "source" in item:
             image = Image.fromarray(
-                cv2.cvtColor(item["source"]["cv2"], cv2.COLOR_BGR2RGB)
+                gpu_cvt_color(item["source"]["cv2"], cv2.COLOR_BGR2RGB)
             )
             image = image.resize(
                 (MAPPER_PREVIEW_MAX_WIDTH, MAPPER_PREVIEW_MAX_HEIGHT), Image.LANCZOS
@@ -1281,7 +1282,7 @@ def refresh_data(map: list):
 
         if "target" in item:
             image = Image.fromarray(
-                cv2.cvtColor(item["target"]["cv2"], cv2.COLOR_BGR2RGB)
+                gpu_cvt_color(item["target"]["cv2"], cv2.COLOR_BGR2RGB)
             )
             image = image.resize(
                 (MAPPER_PREVIEW_MAX_WIDTH, MAPPER_PREVIEW_MAX_HEIGHT), Image.LANCZOS
@@ -1329,7 +1330,7 @@ def update_webcam_source(
             }
 
             image = Image.fromarray(
-                cv2.cvtColor(map[button_num]["source"]["cv2"], cv2.COLOR_BGR2RGB)
+                gpu_cvt_color(map[button_num]["source"]["cv2"], cv2.COLOR_BGR2RGB)
             )
             image = image.resize(
                 (MAPPER_PREVIEW_MAX_WIDTH, MAPPER_PREVIEW_MAX_HEIGHT), Image.LANCZOS
@@ -1381,7 +1382,7 @@ def update_webcam_target(
             }
 
             image = Image.fromarray(
-                cv2.cvtColor(map[button_num]["target"]["cv2"], cv2.COLOR_BGR2RGB)
+                gpu_cvt_color(map[button_num]["target"]["cv2"], cv2.COLOR_BGR2RGB)
             )
             image = image.resize(
                 (MAPPER_PREVIEW_MAX_WIDTH, MAPPER_PREVIEW_MAX_HEIGHT), Image.LANCZOS