Mind the RefGAP: Correcting Reference Attention in Diffusion-Based Visual Editing

Yanan Wang; Shengcai Liao; Guangyi Liu; Xiaodan Liang

cs.CVSep 28, 2026

Mind the RefGAP: Correcting Reference Attention in Diffusion-Based Visual Editing

Authors: Yanan Wang, Shengcai Liao, Guangyi Liu, Xiaodan Liang

Organizations: Mohamed bin Zayed University of Artificial Intelligence · Institute of Foundation Models · United Arab Emirates University

Abstract

Reference-guided diffusion editors struggle to faithfully reproduce user-provided references. We identify a potential bottleneck in diffusion editors: many methods provide limited reference-attention allocation. For example, in LoomVideo, edit-region queries assign less than 1% of their attention mass to the reference. We introduce RefGAP, a training-free correction that determines logit-offset magnitudes online at each layer from the reference-attention mass measured during the forward pass. Positive offsets to reference logits strengthen reference usage by edit-region queries, while negative offsets for keep-region queries limit reference-induced changes outside the edit. Two global coefficients control the correction; they are selected once on validation data from four development diffusion editors and held fixed. Across seven diffusion-based image/video editors, RefGAP improves identity fidelity in head swapping and face swapping. RefGAP achieves a fidelity-preservation trade-off comparable to separately tuned constant edit-side biases, without per-approach strength sweeps. Additional experiments on virtual try-on and background replacement evaluate transfer beyond identity editing.

Figures & tables

Figure 1: RefGAP overview. Attention-guided corrections strengthen reference attention inside the edit and suppress it outside, using two coefficients shared across approaches and tasks.

	LoomVideo	FLUX.2-klein	FLUX.2-klein-base	VACE	Qwen-Image-Edit	OmniGen2
$u_{\mathrm{edit}}$	0.22	0.35	0.22	1.79	0.30	0.24
$u_{\mathrm{keep}}$	0.13	0.10	0.15	0.69	0.35	0.28

Table 1: Token-count-normalized reference attention. Measured using head-swap validation data.

Figure 3: Qualitative results across four tasks. Base vs. Ours on JoyAI-Video-Edit (JoyAI), LoomVideo, and OmniGen2. Scores are shown below, with green borders marking gains $\geq 0.05$ .

Head swap

Face swap

Virtual try-on

Background replacement

Approach

Repl.

\uparrow

\uparrow

Kpt.

\downarrow

LPIPS

\downarrow

Repl.

\uparrow

\uparrow

Retr.

\uparrow

Kpt.

\downarrow

LPIPS

\downarrow

DINO

\uparrow

LPIPS

\downarrow

DINO

\uparrow

LPIPS

\downarrow

JoyAI

\to

99% +31

.507

\to

.644 +.137

.043

\to

.085 +.042

.218

\to

.215

-

.003

\to

97% +37

.282

\to

.591 +.309

\to

88% +49

.057

\to

.105 +.048

.086

\to

.108 +.022

.283

\to

.298 +.015

.099

\to

.101 +.002

.242

\to

.286 +.044

.063

\to

.065 +.002

LoomVideo

\to

92% +13

.608

\to

.665 +.057

.107

\to

.098

-

.009

.281

\to

.275

-

.006

\to

84% +13

.449

\to

.466 +.017

\to

75% +12

.083

\to

.086 +.003

.180

\to

.178

-

.002

.174

\to

.183 +.009

.105

\to

.105 .000

.135

\to

.258 +.123

.051

\to

.058 +.007

Klein

\to

100% +9

.726

\to

.790 +.064

.097

\to

.153 +.056

.222

\to

.217

-

.005

\to

100% +2

.704

\to

.820 +.116

\to

96% +6

.065

\to

.082 +.017

.064

\to

.076 +.012

.363

\to

.372 +.009

.093

\to

.092

-

.001

.629

\to

.630 +.001

.038

\to

.037

-

.001

Klein-base

\to

100% +23

.588

\to

.799 +.211

.169

\to

.236 +.067

.326

\to

.261

-

.065

\to

100% +5

.557

\to

.831 +.274

\to

96% +22

.129

\to

.223 +.094

.173

\to

.179 +.006

.409

\to

.465 +.056

.150

\to

.160 +.010

.632

\to

.651 +.019

.162

\to

.136

-

.026

VACE

—

.405

\to

.729 +.324

.198

\to

.200 +.002

.228

\to

.214

-

.014

—

.541

\to

.818 +.277

\to

98% +11

.187

\to

.193 +.006

.018

\to

.019 +.001

.171

\to

.226 +.055

.056

\to

.056 .000

.008

\to

.014 +.006

.008

\to

.008 .000

Table 2: Quantitative results across seven approaches and four tasks. JoyAI, Klein, Klein-base, and Qwen denote JoyAI-Video-Edit, FLUX.2-klein, FLUX.2-klein-base, and Qwen-Image-Edit, respectively. Cells show Base

\to

Ours, with differences shown below. Mean differences average all seven approaches (six for Repl., which is undefined for VACE). Green and red indicate improvements and degradations, respectively; gray denotes

|\Delta|<0.005

. Differences are computed from the displayed values; unrounded paired estimates are given in table 14 .

Head swap

Face swap

Virtual try-on

Background replacement

Approach

Ref.

Pose

Qual.

Ref.

Pose

Qual.

Ref.

Pose

Qual.

Ref.

Pose

Qual.

JoyAI-Video-Edit

28 /0/ 0

3 /25/ 0

5 /23/ 0

50 /7/ 0

3 /50/ 4

5 /50/ 2

37 /5/ 12

0 /52/ 2

5 /46/ 3

19 /25/ 6

0 /50/ 0

1 /46/ 3

LoomVideo

7 /18/ 2

12 /14/ 2

8 /14/ 6

18 /31/ 3

3 /44/ 5

1 /33/ 18

36 /15/ 0

0 /51/ 0

3 /48/ 0

39 /12/ 0

0 /50/ 1

1 /49/ 1

FLUX.2-klein

20 /5/ 0

0 /25/ 0

7 /18/ 0

33 /16/ 2

0 /48/ 3

1 /48/ 2

1 /11/ 1

1 /12/ 0

0 /12/ 1

0 /10/ 0

FLUX.2-klein-base

24 /1/ 0

0 /25/ 0

3 /22/ 0

45 /8/ 0

2 /45/ 6

7 /44/ 2

20 /16/ 18

0 /53/ 1

1 /51/ 2

15 /40/ 1

0 /56/ 0

10 /45/ 1

VACE

24 /0/ 1

0 /24/ 1

6 /15/ 4

35 /19/ 1

5 /48/ 2

2 /50/ 3

57 /0/ 0

5 /52/ 0

8 /49/ 0

—

Table 3: Human evaluation across four tasks. Cells report judgment counts as RefGAP /tie/ base for reference fidelity (Ref.), pose/expression consistency with the input (Pose), and visual quality (Qual.). Ties include “both succeed” and “both fail”; “—” denotes an unevaluated setting.

	RefGAP				best admissible edit-side constant
Approach	ID $\uparrow$	Kpt. $\downarrow$	LPIPS $\downarrow$	$b$	ID $\uparrow$	Kpt. $\downarrow$	LPIPS $\downarrow$
JoyAI-Video-Edit	.644	.085	.215	1.5	.644	.075	.207
LoomVideo	.665	.098	.275	2.5	.671	.098	.271
FLUX.2-klein	.790	.153	.217	2.0	.794	.171	.229
FLUX.2-klein-base	.799	.236	.261	2.0	.801	.236	.260

Table 4: RefGAP versus approach-specific constant edit-side biases. Constants are selected on validation data under

\Delta\mathrm{LPIPS}_{\mathrm{ne}}\leq 0.005

. Only the edit-side bias differs between settings. Results use the full

1{,}040

-clip head-swap test set; LPIPS is measured over the whole frame.

Appendix figures & tables18 assets

Supplementary material from the paper’s appendix.

Appendix

		Denoising (s/sample)			End-to-end (s/sample)			Peak memory
Approach	Steps	Base	RefGAP	$\Delta$	Base	RefGAP	$\Delta$	Base (GB)	$\Delta$ (MiB)
VACE	30	158.2	159.0	$+0.5\%$	159.4	160.0	$+0.4\%$	26.30	$+0$
LoomVideo	50	106.2	108.9	$+2.5\%$	112.0	116.6	$+4.1\%$	43.95	$+1.5$
Qwen-Image-Edit	20	62.48	66.00	$+5.6\%$	64.0	67.4	$+5.3\%$	57.95	$+0$
OmniGen2	50	21.68	23.39	$+7.9\%$	22.4	24.2	$+8.0\%$	15.36	$+0$
FLUX.2-klein-base	50	16.35	20.34	$+24\%$	17.0	21.2	$+25\%$	15.55	$+0$

Appendix

Table 6: Inference cost of RefGAP on head swapping. Means over five timed samples after one warm-up, Base and RefGAP back to back on the same A100-80GB GPU with the deployed settings. Denoising time excludes encoding, decoding, and saving; end-to-end time excludes model loading. Peak memory is the PyTorch allocator’s maximum allocation. † Unfused reference implementation that materializes the attention matrix in corrected layers; its streaming server reports no per-step timing. ‡ Whole-GPU peak from nvidia-smi .

Approach	$\pi_{\mathcal{R}}$	$\bar{\rho}_{\mathrm{edit}}$	$u_{\mathrm{edit}}$	$\bar{\rho}_{\mathrm{keep}}$	$u_{\mathrm{keep}}$	$B$	$\bar{b}_{\mathrm{edit}}$
JoyAI-Video-Edit	n/a	.0615	—	.0555	—	2.30	2.96
LoomVideo	.031	.0069	0.22	.0040	0.13	2.32	5.91
FLUX.2-klein	.286	.0988	0.35	.0290	0.10	4.48	2.68
FLUX.2-klein-base	.286	.0635	0.22	.0433	0.15	1.88	3.24
VACE	.033	.0590	1.79	.0228	0.69	5.99	3.21
Qwen-Image-Edit	.332	.0990	0.30	.1157	0.35	1.11	2.47

Appendix

Table 7: Full reference-attention diagnostics. Measurements from seven base models on the head-swap validation split.

B

is the peak edit-to-keep reference-mass ratio ( eq. 6 );

\bar{b}_{\mathrm{edit}}

is the mean log-inverse edit mass ( eq. 7 ). JoyAI-Video-Edit’s uniform baseline and normalized masses are omitted because its key composition varies during streaming.

	$\gamma_{\mathrm{keep}}{=}0$		$\gamma_{\mathrm{keep}}{=}0.4$		$\gamma_{\mathrm{keep}}{=}0.8$
$\gamma_{\mathrm{edit}}$	ID $\uparrow$	LPIPS ne $\downarrow$	ID	LPIPS ne	ID	LPIPS ne
JoyAI-Video-Edit ( $B{=}2.3$ ); base ID $.557$ , LPIPS ne $.0663$
all layers $\star$
$0.4$	.689	.0615	.683	.0625	.682	.0618
$0.6$	.705	.0635	.705	.0644	.706	.0641
$0.8$	.548	.0913 ×	.543	.0914 ×	.549	.0917 ×

Appendix

Table 8: Joint calibration of correction coefficients and layer coverage. Head-swap validation results.

\star

marks rule-selected layers; bold marks the chosen configuration. LPIPS ne denotes non-edit LPIPS, and × marks an increase over the base exceeding

0.005

	JoyAI-Video-Edit		LoomVideo		FLUX.2-klein		FLUX.2-klein-base
	all		second half		second half		all
$b$	ID $\uparrow$	LPIPS ne $\downarrow$	ID $\uparrow$	LPIPS ne $\downarrow$	ID $\uparrow$	LPIPS ne $\downarrow$	ID $\uparrow$	LPIPS ne $\downarrow$
base	.557	.0663	.653	.0700	.807	.0435	.590	.1206
0.5	.647	.0641	.661	.0681	.851	.0377	.806	.0667
1.0	.683	.0625	.689	.0676	.883	.0394	.861	.0659
1.5	.703	.0616	.717	.0669	.888	.0410	.880	.0667

Appendix

Table 9: Fixed edit-side sweeps that select each approach’s constant (head-swap validation,

40

clips). × marks

\Delta\mathrm{LPIPS}_{\mathrm{ne}}>0.005

relative to the base; bold marks the best admissible constant, the highest-identity value that satisfies the criterion, which is the constant used in table 4 .

	ID (AdaFace) $\uparrow$	Kpt. $\downarrow$	LPIPS $\downarrow$
DirectSwap	$0.7019\rightarrow 0.7450$	$0.0289\rightarrow 0.0371$	$0.1333\rightarrow 0.1339$

Appendix

Table 10: DirectSwap on the HeadSwapBench test split. Scores are reported as base

\rightarrow

RefGAP using clip-level averaging. ID and Kpt. denote identity similarity and keypoint error, respectively. LPIPS is measured over the whole frame.

	JoyAI-Video-Edit	LoomVideo	FLUX.2-klein	FLUX.2-klein-base	VACE	Qwen-Image-Edit	OmniGen2	DirectSwap
base	0.809	0.816	0.861	0.757	0.786	0.734	0.581	0.894
RefGAP	0.876	0.861	0.887	0.875	0.888	0.802	0.851	0.894
win %	73	73	61	76	96	63	96	49

Appendix

Table 11: Whole-head similarity on HeadSwapBench. Win % is the percentage of paired valid clips on which RefGAP outperforms the base. DirectSwap is included as a task-fine-tuned reference.

	ID-Retr. (%) $\uparrow$		ID-sim $\uparrow$		Kpt. $\downarrow$
input	base	RefGAP	base	RefGAP	base	RefGAP
full frame ( $640\times 480$ )	25.0	50.0	.109	.240	.085	.108
face-centered crop ( $512^{2}$ )	40.0	80.0	.248	.434	.058	.074

Appendix

Table 12: Full-frame versus face-centered inputs for face swapping

Approach	DINO $\uparrow$	LPIPS $\downarrow$
FLUX.2-klein	0.382 $\rightarrow$ 0.391	0.0302 $\rightarrow$ 0.0318
FLUX.2-klein-base	0.407 $\rightarrow$ 0.426	0.0415 $\rightarrow$ 0.0417
Qwen-Image-Edit	0.216 $\rightarrow$ 0.225	0.0158 $\rightarrow$ 0.0160
OmniGen2	0.397 $\rightarrow$ 0.443	0.1339 $\rightarrow$ 0.1395

Appendix

Table 13: Virtual try-on on

100

VITON-HD images. Scores are shown as base

\rightarrow

RefGAP.

Task	Metric	JoyAI	LoomVideo	Klein	Klein-base	VACE	Qwen	OmniGen2
Head swap	Repl. (pp) $\uparrow$	+30.9 [+21.5, +40.1]	+12.2 [+6.6, +17.6]	+8.6 [+5.1, +12.3]	+22.8 [+17.5, +27.4]	—	+16.1 [+10.1, +24.1]	+65.1 [+56.7, +71.4]
	ID $\uparrow$	+.137 [+.111, +.169]	+.058 [+.045, +.069]	+.064 [+.049, +.088]	+.211 [+.176, +.247]	+.323 [+.290, +.362]	+.114 [+.071, +.167]	+.486 [+.449, +.525]
	Kpt. $\downarrow$	+.042 [+.027, +.059]	$-$ .009 [ $-$ .025, +.002]	+.055 [+.046, +.065]	+.068 [+.048, +.096]	+.001 [ $-$ .009, +.012]	+.000 [ $-$ .017, +.015]	+.002 [ $-$ .019, +.021]
	LPIPS $\downarrow$	$-$ .003 [ $-$ .010, +.006]	$-$ .005 [ $-$ .013, +.000]	$-$ .004 [ $-$ .011, +.005]	$-$ .066 [ $-$ .093, $-$ .037]	$-$ .013 [ $-$ .016, $-$ .011]	$-$ .212 [ $-$ .246, $-$ .171]	$-$ .099 [ $-$ .117, $-$ .079]
Face swap	Repl. (pp) $\uparrow$	+36.8 [+31.4, +42.6]	+13.0 [+9.4, +17.1]	+2.2 [+0.8, +4.1]	+5.1 [+2.9, +7.8]	—	+16.7 [+11.7, +21.8]	+20.7 [+16.7, +24.9]
	ID $\uparrow$	+.310 [+.293, +.327]	+.017 [+.003, +.034]	+.116 [+.102, +.131]	+.274 [+.244, +.305]	+.277 [+.262, +.294]	+.169 [+.137, +.202]	+.498 [+.474, +.521]

Appendix

Table 14: Paired differences with pointwise 95% bootstrap confidence intervals. Cells show Ours

-

Base and the corresponding interval from 10,000 task-specific resamples. Green/red indicate intervals excluding zero in the improving/degrading direction; gray indicates intervals containing zero. Colors are determined before rounding. Repl. and Retr. differences are in percentage points. VACE Repl. is undefined and omitted. Sample counts vary by metric and detection availability.

	RefGAP	base	both succeed	both fail	tie %
Reference fidelity	78 (27.9%)	10 (3.6%)	144	48	68.6
Pose / expression	20 (7.1%)	10 (3.6%)	178	72	89.3
Visual quality	0 (0.0%)	22 (7.9%)	246	12	92.1

Appendix

Table 15: Unscreened audit form. Judgment counts from five raters on

56

randomly sampled items (two per approach and task), with no screening for visible differences.

Task	Question	RefGAP	both succeed	both fail	base	total
Head swap	closer to reference	157	23	1	3	184
	pose / expression	27	109	46	3	185
	visual quality	37	131	7	10	185
Face swap	closer to reference	278	89	7	8	382
	pose / expression	19	185	138	40	382
	visual quality	32	312	8	30	382

Appendix

Table 16: Breakdown of tied judgments. The “both succeed” and “both fail” responses are reported separately and sum to the tie counts in table 3 . Entries report the number of rater responses.

Task	Ref.	Pose/expr.	Quality	Ties
Head swap	93.5 [84.7,100]	86.0 [66.0,100]	86.2 [71.9,97.9]	12.6 [3.6,23.4]
Face swap	93.6 [87.2,98.7]	39.7 [23.0,57.0]	56.1 [40.0,72.2]	26.0 [16.9,35.4]
Try-on	85.2 [75.4,93.8]	71.4 [42.9,100]	85.0 [70.0,100]	14.6 [6.8,23.5]
Background	82.2 [68.8,93.8]	40.0 [0.0,80.0]	52.2 [30.0,74.4]	45.7 [32.7,58.6]

Appendix

Table 17: Item-level human preferences pooled across approaches (%). Ref., Pose/expr., and Quality report preference for RefGAP after excluding ties. Ties reports the reference-fidelity tie rate. Brackets show

95\%

item-bootstrap confidence intervals.

Approach	$\Delta$ kpt	$\Delta$ ID	lowest $\Delta$ ID quartile	highest $\Delta$ ID quartile	$\|\Delta\mathrm{ID}\|<0.05$ ( $n$ )
JoyAI-Video-Edit	$+0.048$	$+0.310$	$+0.037$	$+0.064$	$+0.029$ (44)
LoomVideo	$+0.003$	$+0.017$	$-0.010$	$+0.012$	$+0.006$ (350)
FLUX.2-klein	$+0.017$	$+0.116$	$+0.011$	$+0.026$	$+0.012$ (348)
FLUX.2-klein-base	$+0.094$	$+0.274$	$+0.055$	$+0.111$	$+0.052$ (230)
VACE	$+0.006$	$+0.278$	$+0.013$	$+0.002$	$+0.025$ (13)
Qwen-Image-Edit	$+0.021$	$+0.169$	$-0.020$	$+0.068$	$+0.012$ (338)

Appendix

Table 18: Face-swap keypoint change versus identity gain. Per clip,

\Delta

kpt is RefGAP minus base keypoint error against the input subject (lower is better), and

\Delta

ID is RefGAP minus base AdaFace similarity to the reference image given to the model. The quartile columns report the average

\Delta

kpt within the lowest and highest quarters of

\Delta

ID. The last column reports the average over clips whose identity similarity changes by less than

0.05

, with the corresponding count in parentheses.

HeadSwapBench	ID $\uparrow$	kpt $\downarrow$	LPIPS $\downarrow$
FLUX.2-klein (4-step schedule)
base	0.726	0.097	0.222
RefGAP ( $\gamma_{\mathrm{edit}}{=}0.6$ , all steps)	0.790	0.153	0.217
RefGAP, skip step 1	0.753	0.101	0.221
JoyAI-Video-Edit (2-step schedule)
base	0.507	0.043	0.218

Appendix

Table 19: Disabling correction at early denoising steps. Results on

1040

head-swap test clips. “Skip” disables the attention correction at the indicated steps; denoising still proceeds normally. LPIPS is measured over the whole frame. Bold marks the best value within each approach.

			diagnostic clips ( $n=30$ )		all clips ( $n=150$ )
setting	layers	edit-side strength	DINO $\uparrow$	subject LPIPS $\downarrow$	DINO $\uparrow$
base	—	—	.473	.0384	.299
RefGAP (shared setting)	all	online, $\gamma_{\mathrm{edit}}=0.6$	.112	.0242	.042
RefGAP, second half	$30$ – $59$	online, $\gamma_{\mathrm{edit}}=0.6$	.106	.0127	—
fixed edit-side bias	all	constant, $b=2.0$	.040	.0174	.037

Appendix

Table 20: Qwen-Image-Edit background replacement under alternative settings. The diagnostic clips are the first

30

of the

150

-clip list, not selected on outcome.

Figure 4: Seed sensitivity of OmniGen2 head-swap outputs. Columns show the reference, input, base, RefGAP at seed $0$ , a repeat at the same seed, and three alternative seeds. Repeated seeds reproduce the failures, while alternative seeds can recover plausible faces.

Figure 5: Complete seven-approach comparison for the main-paper examples. Each block shows the input, reference, and base/RefGAP outputs, with reference-fidelity scores beneath. Green borders indicate gains $\geq 0.05$ .

Figure 6: Additional qualitative examples across seven approaches and four tasks. The layout and annotations follow fig. 5 .

Explore similar work

Toward 360-Degree Indoor Panorama Editing via Tuning-Free Diffusion Model with Refocusing Cross-Attention
Jun 12, 2026Dinh-Khoi Vo, Nhut-Thanh Le-Hinh, Viet-Tham Huynh +3Diffusion-Based Image Editing Omnidirectional Images
University of Science, Ho Chi Minh, Vietnam · Vietnam National University, Ho Chi Minh, Vietnam · University of Dayton, Ohio, United States
Refinement Is Inherently Editable: Training-Free Prompt-to-Prompt Image Editing with Generative Refinement Network
Sep 17, 2026Yulong Chen, Ziqian Zhang, Haoyu Zhang +4Diffusion-Based Image Editing Image Editing
City University of Hong Kong (Dongguan), Guangdong, China · City University of Hong Kong, Hong Kong, China · Mohamed bin Zayed University of Artificial Intelligence, Masdar, Abu Dhabi
Masked Generative Transformer Is What You Need for Image Editing
May 11, 2026Wei Chow, Linfeng Li, Xian Sun +14Diffusion-Based Image Editing Image Editing
1ByteDance · 2National University of Singapore · 3Duke University +2

Jun 12, 2026, cs.CVJ/K move · Enter open · S save

Toward 360-Degree Indoor Panorama Editing via Tuning-Free Diffusion Model with Refocusing Cross-Attention

Dinh-Khoi Vo, Nhut-Thanh Le-Hinh, Viet-Tham Huynh, Tam V. Nguyen+2

University of Science, Ho Chi Minh, Vietnam · Vietnam National University, Ho Chi Minh, Vietnam · University of Dayton, Ohio, United States

Diffusion-Based Image Editing Omnidirectional Images Cross-Attention Zero-Shot

Open paper AskPDF arXiv

Zero-shot text-guided diffusion has significantly advanced image editing; however, its practical usability remains constrained by three persistent challenges: prompt brittleness that requires meticulous prompt engineering, spillover edits that unintentionally affect non-target regions, and failures on small or cluttered objects caused by limited fine-grained supervision in training data. We propose FocusDiff (Target-Aware Refocusing for Tuning-Free Diffusion Editing), a tuning-free framework for precise and region-specific image manipulation based on refocusing cross-attention. Given a target region obtained through automated segmentation or manual selection, FocusDiff applies selective blurring to non-edit areas to guide attention toward the masked region while accurately transferring the object's identity, structure, and appearance to the edited output. Integrated context-preserving modules further ensure background fidelity and global coherence, enabling accurate edits from simple text prompts in a single pass. We also extend FocusDiff to 360-degree indoor panorama editing and demonstrate its effectiveness within virtual reality environments. Extensive experiments on our localized editing benchmark LIMB, comprising 30 multi-object images and 100 annotated examples including challenging small-object cases, show that FocusDiff outperforms existing zero-shot editors in text-image alignment and background preservation, achieving superior precision, photorealism, and usability. The project page is available at https://vdkhoi20.github.io/FocusDiff.

Abstract

Explore similar work

Toward 360-Degree Indoor Panorama Editing via Tuning-Free Diffusion Model with Refocusing Cross-Attention

Refinement Is Inherently Editable: Training-Free Prompt-to-Prompt Image Editing with Generative Refinement Network

Masked Generative Transformer Is What You Need for Image Editing