工作流 /
图像 /
Bria：引导式背景移除
Bria：引导式背景移除
Bria: Guided Background Removal
使用 Bria 移除图像背景，同时保留文本提示词所描述的内容，从而可以从繁杂场景中抠出单个主体，而不必丢失所有非背景内容。原始像素以完整分辨率返回，并附带新的透明度通道。非常适合产品抠图、人像分离和合成准备。
图像
5 个节点 1 个输入 14 次使用 ComfyUI ≥ 0.42.5 ComfyUI 7 KB
节点构成
MarkdownNote
× 1
LoadImage
× 1
BriaGuidedRemoveBackground
× 1
SaveImage
× 1
ImageCompare
× 1
官方标注的模型
BRIA
工作流 JSON
和上面「下载工作流 JSON」是同一份文件。
1{
2 "id": "97c0d4e0-c7dd-4993-8046-184243cab914",
3 "revision": 0,
4 "last_node_id": 9,
5 "last_link_id": 8,
6 "nodes": [
7 {
8 "id": 1,
9 "type": "MarkdownNote",
10 "pos": [
11 40,
12 40
13 ],
14 "size": [
15 420,
16 520
17 ],
18 "flags": [],
19 "order": 0,
20 "mode": 0,
21 "inputs": [],
22 "outputs": [],
23 "properties": [],
24 "widgets_values": [
25 "# Bria: Guided Background Removal\n\nDescribe what to keep and Bria removes everything else, so a cutout can keep one subject out of a busy scene instead of losing whatever is not background. The original pixels come back at full resolution with a new alpha channel.\n\n## Input Asset\n\n- [bria_guided_person.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/bria_guided_person.png)\n\n## Parameters\n\n- The prompt names what stays: `the man in the blue shirt`, `only the shoes`. A prompt that matches nothing returns a fully transparent image.\n- Naming what to drop as well as what to keep is the most reliable way to get a clean edge around hair and thin clothing.\n- `moderation` on exposes separate visual input and output moderation toggles; off skips them.\n- For a plain cutout with no guidance, Bria Remove Image Background is the simpler node.\n\n## Report Issue\n\nNote: Please update ComfyUI first ([guide](https://docs.comfy.org/installation/update_comfyui)) and prepare required models. Desktop/Cloud updates follow stable releases, so some nightly-supported models may not be available yet.\n\n- Cannot run / runtime errors: [ComfyUI/issues](https://github.com/comfyanonymous/ComfyUI/issues)\n- UI / frontend issues: [ComfyUI_frontend/issues](https://github.com/Comfy-Org/ComfyUI_frontend/issues)\n- Workflow issues: [workflow_templates/issues](https://github.com/Comfy-Org/workflow_templates/issues)\n"
26 ],
27 "widgets_values_named": {
28 "text": "# Bria: Guided Background Removal\n\nDescribe what to keep and Bria removes everything else, so a cutout can keep one subject out of a busy scene instead of losing whatever is not background. The original pixels come back at full resolution with a new alpha channel.\n\n## Input Asset\n\n- [bria_guided_person.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/bria_guided_person.png)\n\n## Parameters\n\n- The prompt names what stays: `the man in the blue shirt`, `only the shoes`. A prompt that matches nothing returns a fully transparent image.\n- Naming what to drop as well as what to keep is the most reliable way to get a clean edge around hair and thin clothing.\n- `moderation` on exposes separate visual input and output moderation toggles; off skips them.\n- For a plain cutout with no guidance, Bria Remove Image Background is the simpler node.\n\n## Report Issue\n\nNote: Please update ComfyUI first ([guide](https://docs.comfy.org/installation/update_comfyui)) and prepare required models. Desktop/Cloud updates follow stable releases, so some nightly-supported models may not be available yet.\n\n- Cannot run / runtime errors: [ComfyUI/issues](https://github.com/comfyanonymous/ComfyUI/issues)\n- UI / frontend issues: [ComfyUI_frontend/issues](https://github.com/Comfy-Org/ComfyUI_frontend/issues)\n- Workflow issues: [workflow_templates/issues](https://github.com/Comfy-Org/workflow_templates/issues)\n"
29 },
30 "color": "#222",
31 "bgcolor": "#000"
32 },
33 {
34 "id": 3,
35 "type": "LoadImage",
36 "pos": [
37 500,
38 40
39 ],
40 "size": [
41 320,
42 370
43 ],
44 "flags": [],
45 "order": 1,
46 "mode": 0,
47 "inputs": [],
48 "outputs": [
49 {
50 "name": "IMAGE",
51 "type": "IMAGE",
52 "links": [
53 1,
54 4
55 ]
56 },
57 {
58 "name": "MASK",
59 "type": "MASK",
60 "links": null
61 }
62 ],
63 "properties": {
64 "Node name for S&R": "LoadImage"
65 },
66 "widgets_values": [
67 "bria_guided_person.png",
68 "image"
69 ],
70 "widgets_values_named": {
71 "image": "bria_guided_person.png",
72 "upload": "image"
73 }
74 },
75 {
76 "id": 2,
77 "type": "BriaGuidedRemoveBackground",
78 "pos": [
79 868,
80 40
81 ],
82 "size": [
83 420,
84 300
85 ],
86 "flags": [],
87 "order": 2,
88 "mode": 0,
89 "inputs": [
90 {
91 "name": "image",
92 "type": "IMAGE",
93 "link": 1
94 }
95 ],
96 "outputs": [
97 {
98 "name": "IMAGE",
99 "type": "IMAGE",
100 "links": [
101 2
102 ]
103 }
104 ],
105 "properties": {
106 "Node name for S&R": "BriaGuidedRemoveBackground"
107 },
108 "widgets_values": [
109 "the archer with the bow and the quiver\n",
110 "true",
111 false,
112 false
113 ],
114 "widgets_values_named": {
115 "prompt": "the archer with the bow and the quiver\n",
116 "moderation": "true",
117 "moderation.visual_input_moderation": false,
118 "moderation.visual_output_moderation": false
119 },
120 "color": "#432",
121 "bgcolor": "#653"
122 },
123 {
124 "id": 4,
125 "type": "SaveImage",
126 "pos": [
127 1336,
128 40
129 ],
130 "size": [
131 420,
132 460
133 ],
134 "flags": [],
135 "order": 3,
136 "mode": 0,
137 "inputs": [
138 {
139 "name": "images",
140 "type": "IMAGE",
141 "link": 2
142 }
143 ],
144 "outputs": [
145 {
146 "name": "images",
147 "type": "IMAGE",
148 "links": [
149 3
150 ]
151 }
152 ],
153 "properties": {
154 "Node name for S&R": "SaveImage"
155 },
156 "widgets_values": [
157 "api_bria_guided_remove_background"
158 ],
159 "widgets_values_named": {
160 "filename_prefix": "api_bria_guided_remove_background"
161 }
162 },
163 {
164 "id": 5,
165 "type": "ImageCompare",
166 "pos": [
167 1804,
168 40
169 ],
170 "size": [
171 440.2221916777981,
172 473.5349069800176
173 ],
174 "flags": [],
175 "order": 4,
176 "mode": 0,
177 "inputs": [
178 {
179 "name": "image_a",
180 "shape": 7,
181 "type": "IMAGE",
182 "link": 4
183 },
184 {
185 "name": "image_b",
186 "shape": 7,
187 "type": "IMAGE",
188 "link": 3
189 }
190 ],
191 "outputs": [],
192 "properties": {
193 "Node name for S&R": "ImageCompare"
194 },
195 "widgets_values": [],
196 "widgets_values_named": []
197 }
198 ],
199 "links": [
200 [
201 2,
202 2,
203 0,
204 4,
205 0,
206 "IMAGE"
207 ],
208 [
209 1,
210 3,
211 0,
212 2,
213 0,
214 "IMAGE"
215 ],
216 [
217 3,
218 4,
219 0,
220 5,
221 1,
222 "IMAGE"
223 ],
224 [
225 4,
226 3,
227 0,
228 5,
229 0,
230 "IMAGE"
231 ]
232 ],
233 "groups": [],
234 "config": [],
235 "extra": {
236 "ds": {
237 "scale": 0.3928808160514848,
238 "offset": [
239 1166.151952092833,
240 843.0554056688183
241 ]
242 },
243 "frontendVersion": "1.57.1",
244 "VHS_latentpreview": false,
245 "VHS_latentpreviewrate": 0,
246 "VHS_MetadataImage": true,
247 "VHS_KeepIntermediate": true
248 },
249 "version": 0.4
250}