Hybrid Generative-Discriminative Object Placement

Siyuan Zhou; Li Niu

cs.CVAug 24, 2026

Hybrid Generative-Discriminative Object Placement

Authors: Siyuan Zhou, Li Niu

Organizations: Shanghai Jiao Tong University

Abstract

As an important operation of image composition, object placement aims to predict the plausible placement (location, scale) for the inserted foreground object. Previous object placement methods can be divided into generative methods and discriminative methods, both of which cannot balance efficiency and effectiveness well. In this work, we propose a semi-generative method in the middle ground between them. In particular, we assign uniformly distributed anchors on the background. Then, we fuse foreground and background features to predict the rationality score for each anchor and predict plausible placement sets for positive anchors. Extensive experiments on the OPA dataset show that our method can strike a good balance between efficiency and effectiveness.

Figures & tables

Figure 1: The comparison among three paradigms for object placement learning. Given a pair of foreground with mask and background, the generative model (a) generates one or multiple plausible placements (center location $(x,y)$ and scale $s$ ) for the foreground. The discriminative model (b) predicts a rationality score map for each foreground scale. $r_{i,s}$ is the rationality score for the $i$ -th center location and scale $s$ . Our semi-generative model (c) predicts a rationality score $r_{i}$ for the $i$ -th anchor and also predicts multiple plausible placements (center offset $(ox_{i},oy_{i})$ and scale $s_{i}$ ) for each positive anchor.

Figure 2: The illustration of our network structure. We use DFM to fuse foreground feature vector $\bm{f}^{f}$ and background feature map $\bm{F}^{b}$ . For each $\bm{f}_{i}^{m}$ in the fused feature map $\bm{F}^{m}$ , we use it to predict rationality score $r_{i}$ and concatenate it with random vector $\bm{z}_{t}$ to produce a placement set $\mathcal{P}_{i}$ . The detailed architecture of our Dynamic Fusion Module (DFM) is shown at the top right corner. We concatenate background feature and foreground feature to produce the offset map for deformable convolution, which can dynamically aggregate multi-scale contextual information.

Figure 3: We suppose that an image is divided into $4\times 4$ cells, corresponding to $16$ anchors (red dot). In the left column, we show the foreground/background and the rationality score map. In the right column, we show one foreground placement, which is associated with the $i$ -th anchor because the foreground center is within its neighborhood (red cell). The location of foreground center can be calculated based on the anchor location $(x_{i},y_{i})$ and center offset $(ox_{i},oy_{i})$ . The foreground placement (yellow bounding box) is represented by the foreground center location $(x_{i}\!+\!ox_{i},y_{i}\!+\!oy_{i})$ and the foreground scale $s_{i}$ .

Method	Plausibility			Diversity	Complexity
Method	User Study $\uparrow$	Acc. $\uparrow$	FID $\downarrow$	LPIPS $\uparrow$	FLOPs(G)	Time(s)
TERSE [ 25 ]	0.104	0.679	46.94	0	37.25	0.040
PlaceNet [ 29 ]	0.109	0.683	36.69	0.160	4.76	0.039
GracoNet [ 32 ]	0.183	0.847	27.75	0.206	43.98	0.044
FOPA [ 19 ]	0.303	0.896	25.91	0.216	321.05	0.258
Ours	0.301	0.890	26.02	0.218	39.49	0.041

Table 1: Quantitative comparison of different methods on OPA dataset.

Figure 4: For each foreground-background pair, we show three composite images generated by GracoNet [ 32 ] , FOPA [ 19 ] , and our method.

Row	Method	Plausibility		Diversity
Row	Method	Acc. $\uparrow$	FID $\downarrow$	LPIPS $\uparrow$
1	w/o DC	0.871	26.85	0.212
2	w/o $\mathcal{L}_{ms}$	0.845	27.67	0.176
3	w/o $\mathcal{L}_{pl}$	0.832	28.83	0.196
4	w/o $\mathcal{L}_{cl}$	0.712	44.86	0.188
5	$1$ st stage	0.869	26.93	0.211

Table 2: Quantitative comparison of different ablated versions of our method on OPA dataset.

Figure 5: The comparison of rationality score map between FOPA [ 19 ] and our method. From left to right, we show the foreground, background, the rationality score map generated by FOPA and our method.

Figure 6: For each pair of foreground and background, we show three generated composite images for the same positive anchor with the highest rationality score. The neighborhood of each positive anchor is marked with red box on the background.

Compositional Image Generation

Compositional image generation studies how to combine objects, attributes, spatial relations, and visual references into coherent, controllable images.

See all 85 papers

presto: Efficient, Training-free, and Open-world Object Placement via Imaginary Search
Aug 21, 2026Weixuan Ding, Shang Liu, Hanyu Pei +1MLLM Reasoning Compositional Image Generation
Wuhan University, Wuhan Hubei Province, China · University of Louisville, Louisville KY, USA
HiddenObjects: Scalable Diffusion-Distilled Spatial Priors for Object Placement
Apr 12, 2026Marco Schouten, Ioannis Siglidis, Serge Belongie +1Spatial Reasoning Diffusion Model Distillation
Technical University of Denmark · Pioneer Centre for AI · University of Copenhagen
OSInsert: Towards High-authenticity and High-fidelity Image Composition
Feb 23, 2026Jingyuan Wang, Li NiuCompositional Image Generation
MoE Key Lab of Artificial Intelligence, Shanghai Jiao Tong University

Aug 21, 2026J/K move · Enter open · S save

presto: Efficient, Training-free, and Open-world Object Placement via Imaginary Search

Weixuan Ding, Shang Liu, Hanyu Pei, Zeyan Liu

Wuhan University, Wuhan Hubei Province, China · University of Louisville, Louisville KY, USA

MLLM Reasoning Compositional Image Generation

Open paper AskPDF arXiv

Object placement is critical in image composition, requiring spatially and semantically coherent positioning of objects within diverse scenes. Existing approaches typically rely on hand-crafted rules or supervised learning on limited datasets, which restricts their generalization and interpretability, especially in open-world scenarios involving novel objects and scenes. In this work, we reformulate open-world object placement as a heuristic search task guided by reasoning from a Multimodal Large Language Model (MLLM). We introduce presto, a zero-shot, training-free framework that operates within an imaginary action space to iteratively refine object position and scale. Our coarse-to-fine search strategy ensures fast convergence, and we evaluate two decision-making variants: Metric-guided Selection and MLLM-as-a-judge. Experiments across multiple benchmarks show that presto~achieves state-of-the-art performance, particularly in previously unseen, open-world settings. Human studies further reveal that the MLLM-as-a-judge variant produces more perceptually coherent placements than metric-driven approaches, highlighting a gap between standard evaluation metrics and human visual judgment.

Abstract

Compositional Image Generation

Explore similar work

presto: Efficient, Training-free, and Open-world Object Placement via Imaginary Search

HiddenObjects: Scalable Diffusion-Distilled Spatial Priors for Object Placement

OSInsert: Towards High-authenticity and High-fidelity Image Composition