Update app.py

app.py

@@ -15,6 +15,7 @@ from vincenty import vincenty
 import json
 from collections import Counter
 import mediapy
+from pymatting import cutout
 
 #from depth_anything.dpt import DepthAnything
 #from depth_anything.util.transform import Resize, NormalizeImage, PrepareForNet
@@ -128,6 +129,7 @@ def make_video(video_path, outdir='./vis_video_depth', encoder='vits', blur_data
         filenames.sort()
     
     # os.makedirs(outdir, exist_ok=True)
+    global masks
     
     for k, filename in enumerate(filenames):
         file_size = os.path.getsize(filename)/1024/1024
@@ -152,7 +154,7 @@ def make_video(video_path, outdir='./vis_video_depth', encoder='vits', blur_data
         #out = cv2.VideoWriter(output_path, cv2.VideoWriter_fourcc(*"avc1"), frame_rate, (output_width, frame_height))
         #fourcc = cv2.VideoWriter_fourcc(*'mp4v')
         #out = cv2.VideoWriter(output_path, fourcc, frame_rate, (output_width, frame_height))
-        global masks
+        
         count = 0
         n = 0
         depth_frames = []
@@ -236,6 +238,22 @@ def make_video(video_path, outdir='./vis_video_depth', encoder='vits', blur_data
 
             cv2.imwrite(f"f{count}_mask.png", depth_gray)
             masks.append(f"f{count}_mask.png")
+            
+            depth_gray_trimap = np.zeros(depth_gray.shape[:2], np.uint8)
+            depth_gray_trimap[depth_gray==128] = 0
+            depth_gray_trimap[depth_gray==129] = 255
+            depth_gray_trimap[depth_gray==131] = 127
+
+            cv2.imwrite(f"f{count}_trimap.png", depth_gray_trimap)
+
+            cutout(
+   # input image path
+   f"f{count}.png",
+   # input trimap path
+   f"f{count}_trimap.png",
+   # output cutout path
+   f"f{count}.png")
+            
             count += 1
 
         final_vid = create_video(orig_frames, frame_rate, "orig")
@@ -942,18 +960,7 @@ with gr.Blocks(css=css, js=js) as demo:
             load_all = gr.Checkbox(label="Load all")
 
           with gr.Group():
-            html = gr.HTML(value="""<label for='zoom'>Zoom</label><input id='zoom' type='range' style='width:256px;height:1em;' value='0.8' min='0.157' max='1.57' step='0.001' oninput='
-              if (!BABYLON.Engine.LastCreatedScene.activeCamera.metadata) {
-                var evt = document.createEvent(\"Event\");
-                evt.initEvent(\"click\", true, false);
-                document.getElementById(\"reset_cam\").dispatchEvent(evt);
-              } 
-              BABYLON.Engine.LastCreatedScene.activeCamera.fov = this.value;
-              this.parentNode.childNodes[2].innerText = BABYLON.Engine.LastCreatedScene.activeCamera.fov;
-
-              document.getElementById(\"model3D\").getElementsByTagName(\"canvas\")[0].style.filter = \"blur(\" + BABYLON.Engine.LastCreatedScene.getNodes()[parseInt(document.getElementById(\"fnum\").getElementsByTagName(\"input\")[0].value)+1].material.pointSize/2.0*Math.sqrt(2.0) + \"px)\";
-            '/><span>0.8</span>""")
-            camera = gr.HTML(value="""<a href='#' id='reset_cam' style='color:white' onclick='
+            camera = gr.HTML(value="""<a href='#' id='reset_cam' style='float:right;color:white' onclick='
               if (!BABYLON.Engine.LastCreatedScene.activeCamera.metadata) {
                 BABYLON.Engine.LastCreatedScene.activeCamera.metadata = { 
                   screenshot: true,
@@ -970,7 +977,18 @@ with gr.Blocks(css=css, js=js) as demo:
               document.getElementById(\"model3D\").getElementsByTagName(\"canvas\")[0].oncontextmenu = function(e){e.preventDefault();}
               document.getElementById(\"model3D\").getElementsByTagName(\"canvas\")[0].ondrag = function(e){e.preventDefault();}
             '>Reset camera</a>""")
-            contrast = gr.HTML(value="""<label for='contrast'>Contrast</label><input id='contrast' type='range' style='width:256px;height:1em;' value='1.0' min='0' max='2' step='0.001' oninput='
+            html = gr.HTML(value="""<label for='zoom' style='width:5em'>Zoom</label><input id='zoom' type='range' style='width:256px;height:1em;' value='0.8' min='0.157' max='1.57' step='0.001' oninput='
+              if (!BABYLON.Engine.LastCreatedScene.activeCamera.metadata) {
+                var evt = document.createEvent(\"Event\");
+                evt.initEvent(\"click\", true, false);
+                document.getElementById(\"reset_cam\").dispatchEvent(evt);
+              } 
+              BABYLON.Engine.LastCreatedScene.activeCamera.fov = this.value;
+              this.parentNode.childNodes[2].innerText = BABYLON.Engine.LastCreatedScene.activeCamera.fov;
+
+              document.getElementById(\"model3D\").getElementsByTagName(\"canvas\")[0].style.filter = \"blur(\" + BABYLON.Engine.LastCreatedScene.getNodes()[parseInt(document.getElementById(\"fnum\").getElementsByTagName(\"input\")[0].value)+1].material.pointSize/2.0*Math.sqrt(2.0) + \"px)\";
+            '/><span>0.8</span>""")
+            contrast = gr.HTML(value="""<label for='contrast' style='width:5em'>Contrast</label><input id='contrast' type='range' style='width:256px;height:1em;' value='1.0' min='0' max='2' step='0.001' oninput='
               if (!BABYLON.Engine.LastCreatedScene.activeCamera.metadata) {
                 var evt = document.createEvent(\"Event\");
                 evt.initEvent(\"click\", true, false);
@@ -979,7 +997,7 @@ with gr.Blocks(css=css, js=js) as demo:
               BABYLON.Engine.LastCreatedScene.activeCamera.metadata.pipeline.imageProcessing.contrast = this.value;
               this.parentNode.childNodes[2].innerText = BABYLON.Engine.LastCreatedScene.activeCamera.metadata.pipeline.imageProcessing.contrast;
             '/><span>1.0</span>""")
-            exposure = gr.HTML(value="""<label for='exposure'>Exposure</label><input id='exposure' type='range' style='width:256px;height:1em;' value='1.0' min='0' max='2' step='0.001' oninput='
+            exposure = gr.HTML(value="""<label for='exposure' style='width:5em'>Exposure</label><input id='exposure' type='range' style='width:256px;height:1em;' value='1.0' min='0' max='2' step='0.001' oninput='
               if (!BABYLON.Engine.LastCreatedScene.activeCamera.metadata) {
                 var evt = document.createEvent(\"Event\");
                 evt.initEvent(\"click\", true, false);