only show refined clothed human

.gitignore

@@ -14,3 +14,5 @@ neural_voxelization_layer/
 pytorch3d/
 force_push.sh
 results/
+gradio_cached_examples/
+gradio_queue.db

app.py

@@ -47,7 +47,7 @@ description = '''
 </th>
 </table>
 
-<h4> The reconstruction + refinement + video take about 80~200 seconds for single image. <span style="color:red"> If ERROR, try "Submit Image" again.</span></h4>
+<h4> The reconstruction + refinement + video take about 200 seconds for single image. <span style="color:red"> If ERROR, try "Submit Image" again.</span></h4>
 
 <details>
 
@@ -94,7 +94,7 @@ def generate_image(seed, psi):
 
 
 random.seed(2022)
-model_types = ['icon-filter', 'pifu', 'pamir']
+model_types = ['ICON', 'PIFu', 'PaMIR']
 examples = [[item, random.choice(model_types)]
             for item in glob.glob('examples/*.png')]
 
@@ -134,15 +134,12 @@ with gr.Blocks() as demo:
                 clear_color=[0.0, 0.0, 0.0, 0.0],  label="SMPL")
             out_smpl_download = gr.File(label="Download SMPL mesh")
             out_smpl_npy_download = gr.File(label="Download SMPL params")
-            out_recon = gr.Model3D(
-                clear_color=[0.0, 0.0, 0.0, 0.0],  label="Recon")
-            out_recon_download = gr.File(label="Download clothed human mesh")
             out_final = gr.Model3D(
                 clear_color=[0.0, 0.0, 0.0, 0.0],  label="Refined Recon")
             out_final_download = gr.File(
                 label="Download refined clothed human mesh")
 
-    out_lst = [out_smpl, out_smpl_download, out_smpl_npy_download, out_recon, out_recon_download,
+    out_lst = [out_smpl, out_smpl_download, out_smpl_npy_download,
                out_final, out_final_download, out_vid, out_vid_download, overlap_inp]
 
     btn_submit.click(fn=generate_model, inputs=[

apps/infer.py

@@ -49,9 +49,14 @@ logging.getLogger("trimesh").setLevel(logging.ERROR)
 def generate_model(in_path, model_type):
 
     torch.cuda.empty_cache()
+    
+    if model_type == 'ICON':
+        model_type = 'icon-filter'
+    else:
+        model_type = model_type.lower()
 
     config_dict = {'loop_smpl': 100,
-                   'loop_cloth': 100,
+                   'loop_cloth': 200,
                    'patience': 5,
                    'out_dir': './results',
                    'hps_type': 'pymaf',
@@ -340,10 +345,6 @@ def generate_model(in_path, model_type):
             os.path.join(config_dict['out_dir'], cfg.name,
                          f"obj/{data['name']}_recon.obj")
         )
-        recon_obj.export(
-            os.path.join(config_dict['out_dir'], cfg.name,
-                         f"obj/{data['name']}_recon.glb")
-        )
 
         # Isotropic Explicit Remeshing for better geometry topology
         verts_refine, faces_refine = remesh(os.path.join(config_dict['out_dir'], cfg.name,
@@ -428,12 +429,8 @@ def generate_model(in_path, model_type):
                 f"{config_dict['out_dir']}/{cfg.name}/obj/{data['name']}_refine.glb")
 
         # always export visualized video regardless of the cloth refinment
-        if final is not None:
-            verts_lst = [verts_pr, final.vertices]
-            faces_lst = [faces_pr, final.faces]
-        else:
-            verts_lst = [verts_pr]
-            faces_lst = [faces_pr]
+        verts_lst = [smpl_obj.vertices, final.vertices]
+        faces_lst = [smpl_obj.faces, final.faces]
 
         # self-rotated video
         dataset.render.load_meshes(
@@ -447,8 +444,6 @@ def generate_model(in_path, model_type):
     smpl_obj_path = f"{config_dict['out_dir']}/{cfg.name}/obj/{data['name']}_smpl.obj"
     smpl_glb_path = f"{config_dict['out_dir']}/{cfg.name}/obj/{data['name']}_smpl.glb"
     smpl_npy_path = f"{config_dict['out_dir']}/{cfg.name}/obj/{data['name']}_smpl.npy"
-    recon_obj_path = f"{config_dict['out_dir']}/{cfg.name}/obj/{data['name']}_recon.obj"
-    recon_glb_path = f"{config_dict['out_dir']}/{cfg.name}/obj/{data['name']}_recon.glb"
     refine_obj_path = f"{config_dict['out_dir']}/{cfg.name}/obj/{data['name']}_refine.obj"
     refine_glb_path = f"{config_dict['out_dir']}/{cfg.name}/obj/{data['name']}_refine.glb"
 
@@ -464,8 +459,6 @@ def generate_model(in_path, model_type):
     gc.collect()
     torch.cuda.empty_cache()
 
-    return [smpl_glb_path, smpl_obj_path,
-            smpl_npy_path,
-            recon_glb_path, recon_obj_path,
+    return [smpl_glb_path, smpl_obj_path,smpl_npy_path,
             refine_glb_path, refine_obj_path,
             video_path, video_path, overlap_path]