Split guide protrusions raise slit contact area in rollable displays, improving curved guidance and impact resistance.
Uniform shielding across fold and non-fold areas uses a bridging member to join EMI layers and the digitizer with lower structural complexity.
Local input reuse across tiled MAC circuits cuts memory access delays, bandwidth demand, and power in large AI model acceleration.
Grouping same-value word-line inputs cuts CiM MAC input cycles and power, while readout calibration restores accurate output sums.
Textured cover glass with transparent deformation objects reduces bending stress while preserving screen protection and a clear visual appearance.
Preformed grooves in the cover glass confine optically clear resin during lamination, reducing layer separation risk and improving display reliability.
A always-on control block saves settings, powers down the functional block, and restores it for low-power touch display wake-up.
A pre-trained model sets playback volume from audio, user-group, and scenario inputs to avoid repeated manual adjustment and fixed presets.
A lattice support plate and graded adhesive zones ease folding stress while preserving adhesion and structural integrity in flexible displays.
A slotted heat sink with capillary coolant channels and vapor-chamber transfer cuts thermal resistance and improves CPU temperature uniformity.
A partition wall, support plate, and waterproof member seal frame-cover gaps to block foreign substances without sacrificing assembly ease.
A shared product-sum, shift, and sign-handling path supports multiple convolution precisions while reducing circuit area in AI inference.
Icons mirror the real display layout and distinguish embedded from external screens, making dragged window placement easier and clearer.
Loads only the active PQC algorithm constants into accelerator memory, cutting storage needs and power while preserving flexibility.
A centralized OS controller unifies vendor-specific interfaces to allocate power across components and improve system-wide efficiency.
Power headroom monitoring lets a PLD switch between turbo processing and deep sleep to finish tasks faster while cutting energy use.
Sensor feedback detects panel movement and tilt during rolling, allowing controller adjustment for stable winding and unwinding.
An acceleration-triggered disconnect protects a rollable display drive assembly from drop impact while preserving smooth extension and retraction.
A movable microphone layout across dual housings preserves sound pickup during roll and unroll states while enabling display state detection.
Capacitive-coupled IMC SRAM macros and a custom ISA cut off-chip data movement, lowering DNN inference latency and energy use.
Shared-capacitor sub-cells cut transistor count and area in mixed-signal in-memory MAC arrays while reducing charge-sharing errors and power.
Automatic blade translation moves a flexible display between retracted, peek, and extended positions to enlarge screen area without increasing housing bulk.
Low-power sensors and user behavior modeling trigger always-on vision only when needed, cutting false positives, wake delay, and battery drain.
Fusing accelerometer and gyroscope data with hinge-axis constraints improves hinge angle tracking during shaking, large acceleration, and gravity alignment.
Dynamic latency and idle-transition tracking lets serverless clusters reallocate cores and tune CPU frequency to meet SLOs with lower power use.
Liquid-cooled cold plates remove heat from processor, memory, and drive components while cutting fan noise, energy use, and footprint.
Blank-period reset and sequential scan driving extend under-display fingerprint sensing time and improve optical read accuracy.
Dual auxetic layers with negative Poisson's ratio absorb folding stress and impact, helping preserve display surface quality and durability.
Bypass lines routed around a display hole preserve data-line connectivity while saving fan-out space and reducing parasitic capacitance and touch interference.
Fluid temperature sensing adjusts motor speed and flow between server heating elements to limit post-heat-exchange cooling loss.
A screen management hub creates host units for each display, enabling centralized cross-device content control without OS-specific projection limits.
Dynamic power mode switching lets MR positioning edge devices run imaging and ML tasks with lower heat and less throttling.
A metal-plate bending guide controls chip-on-film curvature to cut bezel protrusion while buffering impact and preventing damage.
A hinged optical plate and support structure let an aerial image interface fold flat for transport while holding a stable display angle in use.
Fin channels, drainage paths, and hydrophobic surfaces divert cleaning fluid away from vent apertures to protect internal electronics.
A hybrid digital-analog CNN circuit cuts multiplications, power use, and circuit area by splitting convolution and fully connected processing.
Multiple outer product instructions split vectors into sub-vectors to fill 2D array storage more efficiently and raise processing throughput.
EC telemetry lets a pre-boot app adjust display brightness and power states to conserve battery life before the OS loads.
Program workload hints trigger sub-ambient processor cooling before heat rises, cutting thermal lag and avoiding unnecessary cooling power.
Segmented counter circuits cache LSBs and update MSBs in RAM to cut ripple-carry delay and chip area in high-capacity counting.
A seated center bar stabilizes and guides the FPCB bending area, improving foldable device integrity across folded and unfolded states.
A cover sliding along the display edge protects exposed components while freeing internal space and allowing denser, more independent placement.
Microfluidic and nanofluidic channels with nanocoatings and nanoparticle coolant cut flow resistance and handle high chiplet heat loads.
Inclined bars and localized etching form curved recess boundaries in foldable UTG windows, improving impact resistance and surface quality.
A unified media browsing interface automates location-based content access and privacy controls while keeping content discovery efficient.
A U-shaped sound pickup path cuts microphone stack height in thin foldable electronics while preserving sound pickup performance.
Zero-detection in compute-in-memory circuits disables memory and multiply steps when operands are zero, cutting power without hurting throughput.
UV-cured resin patterns place protective film only in non-display areas, cutting material use and reducing transfer-induced scratches.
A sealing member bridges the inner mold, cover layer, and display panel to narrow bezel width while preserving durability and assembly efficiency.
A grooved outer protective layer and inner bend protection help prevent edge impact damage and resin defects in bent display panels.