update

demo/assets/color_list_CN.csv

@@ -3,4 +3,5 @@ Name,Hex
 白色,ffffff
 红色,d74532
 黑色,000000
-深蓝色,4b6190
+深蓝色,4b6190
+浅灰色,f2f0f0

demo/assets/color_list_EN.csv

@@ -3,4 +3,5 @@ Blue,628bce
 White,ffffff  
 Red,d74532  
 Black,000000  
-Dark Blue,4b6190
+Dark Blue,4b6190
+Light Gray,f2f0f0

demo/locals.py

@@ -252,6 +252,22 @@ LOCALES = {
             "label": "高清照",
         },
     },
+    "standard_photo_png": {
+        "en": {
+            "label": "Matting Standard photo",
+        },
+        "zh": {
+            "label": "透明标准照",
+        },
+    },
+    "hd_photo_png": {
+        "en": {
+            "label": "Matting HD photo",
+        },
+        "zh": {
+            "label": "透明高清照",
+        },
+    },
     "layout_photo": {
         "en": {
             "label": "Layout photo",
@@ -268,4 +284,12 @@ LOCALES = {
             "label": "下载调整 KB 大小后的照片",
         },
     },
+    "matting_image": {
+        "en": {
+            "label": "Matting image",
+        },
+        "zh": {
+            "label": "抠图图像",
+        },
+    },
 }

demo/processor.py

@@ -137,6 +137,7 @@ class IDPhotoProcessor:
             return [
                 gr.update(value=None),  # img_output_standard
                 gr.update(value=None),  # img_output_standard_hd
+                gr.update(value=None),  # img_output_standard_hd_png
                 gr.update(visible=False),  # img_output_layout
                 gr.update(  # notification
                     value=LOCALES["notification"][language]["face_error"],
@@ -147,6 +148,10 @@ class IDPhotoProcessor:
 
         else:
             (result_image_standard, result_image_hd, _, _) = result
+
+            result_image_standard_png = np.uint8(result_image_standard)
+            result_image_hd_png = np.uint8(result_image_hd)
+
             if (
                 idphoto_json["render_mode"]
                 == LOCALES["render_mode"][language]["choices"][0]
@@ -272,6 +277,8 @@ class IDPhotoProcessor:
                 return [
                     result_image_standard,  # img_output_standard
                     result_image_hd,  # img_output_standard_hd
+                    result_image_standard_png,  # img_output_standard_hd_png
+                    result_image_hd_png,  # img_output_standard_hd_png
                     result_layout_image,  # img_output_layout
                     gr.update(visible=False),  # notification
                     gr.update(visible=True, value=output_image_path),  # file_download
@@ -280,6 +287,8 @@ class IDPhotoProcessor:
                 return [
                     result_image_standard,  # img_output_standard
                     result_image_hd,  # img_output_standard_hd
+                    result_image_standard_png,  # img_output_standard_hd_png
+                    result_image_hd_png,  # img_output_standard_hd_png
                     result_layout_image,  # img_output_layout
                     gr.update(visible=False),  # notification
                     gr.update(visible=False),  # file_download

demo/ui.py

@@ -16,35 +16,16 @@ def create_ui(
 ):
     DEFAULT_LANG = "zh"
     DEFAULT_HUMAN_MATTING_MODEL = "modnet_photographic_portrait_matting"
-    DEFAULT_FACE_DETECT_MODEL = "mtcnn"
+    DEFAULT_FACE_DETECT_MODEL = "retinaface-resnet50"
 
     if DEFAULT_HUMAN_MATTING_MODEL in human_matting_models:
         human_matting_models.remove(DEFAULT_HUMAN_MATTING_MODEL)
         human_matting_models.insert(0, DEFAULT_HUMAN_MATTING_MODEL)
 
-    css = """
-        #col-left {
-            margin: 0 auto;
-            max-width: 430px;
-        }
-        #col-mid {
-            margin: 0 auto;
-            max-width: 430px;
-        }
-        #col-right {
-            margin: 0 auto;
-            max-width: 430px;
-        }
-        #col-showcase {
-            margin: 0 auto;
-            max-width: 1100px;
-        }
-        #button {
-            color: blue;
-        }
-        """
-
-    demo = gr.Blocks(title="HivisionIDPhotos", css=css)
+    if DEFAULT_FACE_DETECT_MODEL not in face_detect_models:
+        DEFAULT_FACE_DETECT_MODEL = "mtcnn"
+
+    demo = gr.Blocks(title="HivisionIDPhotos")
 
     with demo:
         gr.HTML(load_description(os.path.join(root_dir, "assets/title.md")))
@@ -237,7 +218,9 @@ def create_ui(
                         ],
                     )
 
-                img_but = gr.Button(LOCALES["button"][DEFAULT_LANG]["label"])
+                img_but = gr.Button(
+                    LOCALES["button"][DEFAULT_LANG]["label"], elem_id="btn"
+                )
 
                 example_images = gr.Examples(
                     inputs=[img_input],
@@ -267,15 +250,34 @@ def create_ui(
                         height=350,
                         format="jpeg",
                     )
+
                 img_output_layout = gr.Image(
                     label=LOCALES["layout_photo"][DEFAULT_LANG]["label"],
                     height=350,
                     format="jpeg",
                 )
+
                 file_download = gr.File(
                     label=LOCALES["download"][DEFAULT_LANG]["label"], visible=False
                 )
 
+                with gr.Accordion(
+                    LOCALES["matting_image"][DEFAULT_LANG]["label"], open=False
+                ) as matting_image_accordion:
+                    with gr.Row():
+                        img_output_standard_png = gr.Image(
+                            label=LOCALES["standard_photo_png"][DEFAULT_LANG]["label"],
+                            height=350,
+                            format="png",
+                            elem_id="standard_photo_png",
+                        )
+                        img_output_standard_hd_png = gr.Image(
+                            label=LOCALES["hd_photo_png"][DEFAULT_LANG]["label"],
+                            height=350,
+                            format="png",
+                            elem_id="hd_photo_png",
+                        )
+
             # ---------------- 设置隐藏/显示组件 ----------------
             def change_language(language):
                 return {
@@ -323,6 +325,12 @@ def create_ui(
                     img_output_standard_hd: gr.update(
                         label=LOCALES["hd_photo"][language]["label"]
                     ),
+                    img_output_standard_png: gr.update(
+                        label=LOCALES["standard_photo_png"][language]["label"]
+                    ),
+                    img_output_standard_hd_png: gr.update(
+                        label=LOCALES["hd_photo_png"][language]["label"]
+                    ),
                     img_output_layout: gr.update(
                         label=LOCALES["layout_photo"][language]["label"]
                     ),
@@ -368,6 +376,9 @@ def create_ui(
                         choices=LOCALES["watermark_switch"][language]["choices"],
                         value=LOCALES["watermark_switch"][language]["choices"][0],
                     ),
+                    matting_image_accordion: gr.update(
+                        label=LOCALES["matting_image"][language]["label"]
+                    ),
                 }
 
             def change_color(colors):
@@ -422,6 +433,8 @@ def create_ui(
                     notification,
                     img_output_standard,
                     img_output_standard_hd,
+                    img_output_standard_png,
+                    img_output_standard_hd_png,
                     img_output_layout,
                     file_download,
                     head_measure_ratio_option,
@@ -436,6 +449,7 @@ def create_ui(
                     watermark_text_angle,
                     watermark_text_space,
                     watermark_options,
+                    matting_image_accordion,
                 ],
             )
 
@@ -484,6 +498,8 @@ def create_ui(
                 outputs=[
                     img_output_standard,
                     img_output_standard_hd,
+                    img_output_standard_png,
+                    img_output_standard_hd_png,
                     img_output_layout,
                     notification,
                     file_download,

hivision/creator/retinaface/inference.py

@@ -3,7 +3,6 @@ import cv2
 import onnxruntime as ort
 from hivision.creator.retinaface.box_utils import decode, decode_landm
 from hivision.creator.retinaface.prior_box import PriorBox
-import argparse
 
 
 def py_cpu_nms(dets, thresh):
@@ -37,31 +36,15 @@ def py_cpu_nms(dets, thresh):
     return keep
 
 
-parser = argparse.ArgumentParser(description="Retinaface")
-
-parser.add_argument(
-    "--network", default="resnet50", help="Backbone network mobile0.25 or resnet50"
-)
-parser.add_argument(
-    "--cpu", action="store_true", default=False, help="Use cpu inference"
-)
-parser.add_argument(
-    "--confidence_threshold", default=0.8, type=float, help="confidence_threshold"
-)
-parser.add_argument("--top_k", default=5000, type=int, help="top_k")
-parser.add_argument("--nms_threshold", default=0.2, type=float, help="nms_threshold")
-parser.add_argument("--keep_top_k", default=750, type=int, help="keep_top_k")
-parser.add_argument(
-    "-s",
-    "--save_image",
-    action="store_true",
-    default=True,
-    help="show detection results",
-)
-parser.add_argument(
-    "--vis_thres", default=0.6, type=float, help="visualization_threshold"
-)
-args = parser.parse_args()
+# 替换掉 argparse 的部分，直接使用普通变量
+network = "resnet50"
+use_cpu = False
+confidence_threshold = 0.8
+top_k = 5000
+nms_threshold = 0.2
+keep_top_k = 750
+save_image = True
+vis_thres = 0.6
 
 
 def load_model_ort(model_path):
@@ -112,7 +95,6 @@ def retinaface_detect_faces(image, model_path: str, sess=None):
     inputs = {"input": img}
     loc, conf, landms = retinaface.run(None, inputs)
 
-    # tic = time.time()
     priorbox = PriorBox(cfg, image_size=(im_height, im_width))
     priors = priorbox.forward()
 
@@ -141,30 +123,28 @@ def retinaface_detect_faces(image, model_path: str, sess=None):
     landms = landms * scale1 / resize
 
     # ignore low scores
-    inds = np.where(scores > args.confidence_threshold)[0]
+    inds = np.where(scores > confidence_threshold)[0]
     boxes = boxes[inds]
     landms = landms[inds]
     scores = scores[inds]
 
     # keep top-K before NMS
-    order = scores.argsort()[::-1][: args.top_k]
+    order = scores.argsort()[::-1][:top_k]
     boxes = boxes[order]
     landms = landms[order]
     scores = scores[order]
 
     # do NMS
     dets = np.hstack((boxes, scores[:, np.newaxis])).astype(np.float32, copy=False)
-    keep = py_cpu_nms(dets, args.nms_threshold)
-    # keep = nms(dets, args.nms_threshold,force_cpu=args.cpu)
+    keep = py_cpu_nms(dets, nms_threshold)
     dets = dets[keep, :]
     landms = landms[keep]
 
     # keep top-K faster NMS
-    dets = dets[: args.keep_top_k, :]
-    landms = landms[: args.keep_top_k, :]
+    dets = dets[:keep_top_k, :]
+    landms = landms[:keep_top_k, :]
 
     dets = np.concatenate((dets, landms), axis=1)
-    # print("post processing time: {:.4f}s".format(time.time() - tic))
 
     return dets, retinaface